Allow ReflectionPad to accept 0-dim batch sizes. #39231

v0dro · 2020-05-29T08:43:30Z

Allows ReflectionPad 1D and 2D to accept 0-dim batch sizes.

Related to issues:

dr-ci · 2020-05-29T08:53:51Z

💊 CI failures summary and remediations

As of commit c2020a3 (more details on the Dr. CI page):

7/7 failures introduced in this PR

🕵️ 7 new failures recognized by patterns

The following CI failures do not appear to be due to upstream breakages:

pytorch_windows_vs2019_py36_cuda10.1_test1 (1/7)

Step: "Test" (full log | diagnosis details | 🔁 rerun)

RuntimeError: test_nn failed!

  File "C:\Jenkins\Miniconda3\lib\site-packages\scipy\_distributor_init.py", line 61, in <module> 
    WinDLL(os.path.abspath(filename)) 
  File "C:\Jenkins\Miniconda3\lib\ctypes\__init__.py", line 348, in __init__ 
    self._handle = _dlopen(self._name, mode) 
OSError: [WinError 126] The specified module could not be found 
Traceback (most recent call last): 
  File "run_test.py", line 721, in <module> 
    main() 
  File "run_test.py", line 714, in main 
    raise RuntimeError(message) 
RuntimeError: test_nn failed! 
 
(base) circleci@PACKER-5EE89583 C:\Users\circleci\project\test>if ERRORLEVEL 1 exit /b 1  
+ cleanup
+ retcode=1
+ set +x

pytorch_linux_xenial_py3_6_gcc5_4_build (2/7)

Step: "(Optional) Merge target branch" (full log | diagnosis details | 🔁 rerun)

Automatic merge failed; fix conflicts and then commit the result.

CONFLICT (add/add): Merge conflict in .circleci/scripts/binary_ios_upload.sh 
Auto-merging .circleci/scripts/binary_ios_upload.sh 
CONFLICT (add/add): Merge conflict in .circleci/docker/common/install_base.sh 
Auto-merging .circleci/docker/common/install_base.sh 
CONFLICT (add/add): Merge conflict in .circleci/docker/build.sh 
Auto-merging .circleci/docker/build.sh 
CONFLICT (add/add): Merge conflict in .circleci/config.yml 
Auto-merging .circleci/config.yml 
CONFLICT (add/add): Merge conflict in .circleci/cimodel/data/simple/docker_definitions.py 
Auto-merging .circleci/cimodel/data/simple/docker_definitions.py 
Automatic merge failed; fix conflicts and then commit the result.

pytorch_windows_vs2019_py36_cuda10.1_on_cpu_test1 (3/7)

Step: "Test" (full log | diagnosis details | 🔁 rerun)

RuntimeError: test_nn failed!

  File "C:\Jenkins\Miniconda3\lib\site-packages\scipy\_distributor_init.py", line 61, in <module> 
    WinDLL(os.path.abspath(filename)) 
  File "C:\Jenkins\Miniconda3\lib\ctypes\__init__.py", line 348, in __init__ 
    self._handle = _dlopen(self._name, mode) 
OSError: [WinError 126] The specified module could not be found 
Traceback (most recent call last): 
  File "run_test.py", line 721, in <module> 
    main() 
  File "run_test.py", line 714, in main 
    raise RuntimeError(message) 
RuntimeError: test_nn failed! 
 
(base) circleci@PACKER-5EE89590 C:\Users\circleci\project\test>if ERRORLEVEL 1 exit /b 1  
+ cleanup
+ retcode=1
+ set +x

pytorch-linux-xenial-py3-clang5-android-ndk-r19c-gradle-build-x86_32 (4/7)

Step: "pytorch android gradle build only x86_32 (for PR)" (full log | diagnosis details | 🔁 rerun)

Jun 24 06:31:53 ./.circleci/scripts/build_android_gradle.sh: line 73: /opt/gradle/gradle-6.5/bin/gradle: No such file or directory

Jun 24 06:31:53 ANDROID_NDK_HOME=/opt/ndk 
Jun 24 06:31:53 INSTALLED_DB= 
Jun 24 06:31:53 _=/usr/bin/env 
Jun 24 06:31:53 + export GRADLE_LOCAL_PROPERTIES=/var/lib/jenkins/workspace/android/local.properties 
Jun 24 06:31:53 + GRADLE_LOCAL_PROPERTIES=/var/lib/jenkins/workspace/android/local.properties 
Jun 24 06:31:53 + rm -f /var/lib/jenkins/workspace/android/local.properties 
Jun 24 06:31:53 + echo sdk.dir=/opt/android/sdk 
Jun 24 06:31:53 + echo ndk.dir=/opt/ndk 
Jun 24 06:31:53 + echo cmake.dir=/usr/local 
Jun 24 06:31:53 + /opt/gradle/gradle-6.5/bin/gradle -p android assembleRelease --debug --stacktrace -PABI_FILTERS=x86 --offline 
Jun 24 06:31:53 ./.circleci/scripts/build_android_gradle.sh: line 73: /opt/gradle/gradle-6.5/bin/gradle: No such file or directory

pytorch_windows_vs2019_py36_cpu_test1 (5/7)

Step: "Test" (full log | diagnosis details | 🔁 rerun)

RuntimeError: test_nn failed!

  File "C:\Jenkins\Miniconda3\lib\site-packages\scipy\_distributor_init.py", line 61, in <module> 
    WinDLL(os.path.abspath(filename)) 
  File "C:\Jenkins\Miniconda3\lib\ctypes\__init__.py", line 348, in __init__ 
    self._handle = _dlopen(self._name, mode) 
OSError: [WinError 126] The specified module could not be found 
Traceback (most recent call last): 
  File "run_test.py", line 721, in <module> 
    main() 
  File "run_test.py", line 714, in main 
    raise RuntimeError(message) 
RuntimeError: test_nn failed! 
 
(base) circleci@PACKER-5EE89590 C:\Users\circleci\project\test>if ERRORLEVEL 1 exit /b 1  
+ cleanup
+ retcode=1
+ set +x

pytorch_windows_vs2019_py36_cuda10.1_test2 (6/7)

Step: "Test" (full log | diagnosis details | 🔁 rerun)

RuntimeError: test_cuda failed!

  File "C:\Jenkins\Miniconda3\lib\site-packages\scipy\_distributor_init.py", line 61, in <module> 
    WinDLL(os.path.abspath(filename)) 
  File "C:\Jenkins\Miniconda3\lib\ctypes\__init__.py", line 348, in __init__ 
    self._handle = _dlopen(self._name, mode) 
OSError: [WinError 126] The specified module could not be found 
Traceback (most recent call last): 
  File "run_test.py", line 721, in <module> 
    main() 
  File "run_test.py", line 714, in main 
    raise RuntimeError(message) 
RuntimeError: test_cuda failed! 
 
(base) circleci@PACKER-5EE89583 C:\Users\circleci\project\test>if ERRORLEVEL 1 exit /b 1  
+ cleanup
+ retcode=1
+ set +x

pytorch_xla_linux_bionic_py3_6_clang9_build (7/7)

Step: "(Optional) Merge target branch" (full log | diagnosis details | 🔁 rerun)

Automatic merge failed; fix conflicts and then commit the result.

CONFLICT (add/add): Merge conflict in .circleci/scripts/binary_ios_upload.sh 
Auto-merging .circleci/scripts/binary_ios_upload.sh 
CONFLICT (add/add): Merge conflict in .circleci/docker/common/install_base.sh 
Auto-merging .circleci/docker/common/install_base.sh 
CONFLICT (add/add): Merge conflict in .circleci/docker/build.sh 
Auto-merging .circleci/docker/build.sh 
CONFLICT (add/add): Merge conflict in .circleci/config.yml 
Auto-merging .circleci/config.yml 
CONFLICT (add/add): Merge conflict in .circleci/cimodel/data/simple/docker_definitions.py 
Auto-merging .circleci/cimodel/data/simple/docker_definitions.py 
Automatic merge failed; fix conflicts and then commit the result.

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group.

See how this bot performed.

This comment has been revised 21 times.

mruberry · 2020-06-02T05:09:30Z

CUDA test failures are real. Do you know what the issue is?

v0dro · 2020-06-05T05:15:23Z

@mruberry problems have been resolved. Could you please review?

mruberry · 2020-06-05T05:32:25Z

test/test_nn.py


+    def test_ReflectionPad_empty(self, device):
+        for mod, inp in [
+                (torch.nn.ReflectionPad1d(2), torch.randn(0, 3, 10, device=device)),


I don't think you're covering all the cases? Don't you need to check ReflectionPad1D with a 2D and 3D input with a zero batch size plus ReflectionPad2D with a 3D and 4D input with a zero batch size?

You should also add test cases that fail, where the non-batch dimension is zero and you assert you hit the appropriate error (self.assertRaisesRegex(...)), and a check for backward when the batch dim is zero.

Ah yes good point. Will add them thank you.

The checks for 2D/3D input are done in the python code in functional.py and the ReflectionPad1D does not get called unless the input is specifically dim=3. Same goes for 2D reflection pad (with dim=4). Therefore such tests are not needed.

However I have added tests for non-batch dimension being zero. The backward is tested in the _test_module_empty_input.

mruberry · 2020-06-05T05:33:35Z

@mruberry problems have been resolved. Could you please review?

Looks pretty good and clean, but you need to add more test cases to validate all the new behavior.

v0dro · 2020-06-20T01:13:34Z

@mruberry can you please review this PR again?

mruberry

Thanks for the follow-up, v0dro! I added a decorator to the test that should resolve the XLA build failure. If the tests pass this should be good to go!

facebook-github-bot

@mruberry has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.

facebook-github-bot · 2020-06-25T06:12:53Z

@mruberry merged this pull request in cf8a9b5.

mruberry · 2020-06-25T17:13:15Z

This was successfully merged. Thanks v0dro!

v0dro added 4 commits May 29, 2020 14:58

reflection pad tests

32e7f43

reflection pad update

04d4003

reflection pad 2d update

7dbad25

update CUDA version of reflection pad

afa5af2

pytorchbot added the open source label May 29, 2020

reflectionPad conditions

606b0f7

gchanan requested review from albanD and mruberry June 1, 2020 14:58

gchanan added the triaged This issue has been looked at a team member, and triaged and prioritized into an appropriate module label Jun 1, 2020

reflection pad update

39446bd

mruberry reviewed Jun 5, 2020

View reviewed changes

v0dro added 3 commits June 19, 2020 06:59

Merge branch 'master' of github.com:pytorch/pytorch into reflection-0dim

3195943

update test

9c0c366

ReflectionPad test update

32e07b7

Adds onlyOnCPUAndCUDA decorator

c2020a3

mruberry self-requested a review June 24, 2020 05:39

mruberry approved these changes Jun 24, 2020

View reviewed changes

facebook-github-bot reviewed Jun 24, 2020

View reviewed changes

facebook-github-bot closed this in cf8a9b5 Jun 25, 2020

facebook-github-bot added the merged label Jun 25, 2020

mruberry added the Merged label Oct 28, 2020

Allow ReflectionPad to accept 0-dim batch sizes. #39231

Allow ReflectionPad to accept 0-dim batch sizes. #39231

Uh oh!

Conversation

v0dro commented May 29, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dr-ci bot commented May 29, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

💊 CI failures summary and remediations

🕵️ 7 new failures recognized by patterns

pytorch_windows_vs2019_py36_cuda10.1_test1 (1/7)

pytorch_linux_xenial_py3_6_gcc5_4_build (2/7)

pytorch_windows_vs2019_py36_cuda10.1_on_cpu_test1 (3/7)

pytorch-linux-xenial-py3-clang5-android-ndk-r19c-gradle-build-x86_32 (4/7)

pytorch_windows_vs2019_py36_cpu_test1 (5/7)

pytorch_windows_vs2019_py36_cuda10.1_test2 (6/7)

pytorch_xla_linux_bionic_py3_6_clang9_build (7/7)

Uh oh!

mruberry commented Jun 2, 2020

Uh oh!

v0dro commented Jun 5, 2020

Uh oh!

mruberry Jun 5, 2020

Choose a reason for hiding this comment

Uh oh!

v0dro Jun 5, 2020

Choose a reason for hiding this comment

Uh oh!

v0dro Jun 19, 2020

Choose a reason for hiding this comment

Uh oh!

mruberry commented Jun 5, 2020

Uh oh!

v0dro commented Jun 20, 2020

Uh oh!

mruberry left a comment

Choose a reason for hiding this comment

Uh oh!

facebook-github-bot left a comment

Choose a reason for hiding this comment

Uh oh!

facebook-github-bot commented Jun 25, 2020

Uh oh!

mruberry commented Jun 25, 2020

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

v0dro commented May 29, 2020 •

edited

Loading

dr-ci bot commented May 29, 2020 •

edited

Loading