Skip to content

Conversation

@Xia-Weiwen
Copy link
Collaborator

@Xia-Weiwen Xia-Weiwen commented Jan 3, 2023

Summary
Fixes #91551
aten.pixel_shuffle.default is not registered for meta and it always generates contiguous (channels-first) layout of outputs. It can be reproduced by torch.compile (as described in the issue #91551) and running in FakeTensorMode.

Test plan
python test/inductor/test_torchinductor.py -k test_pixel_shuffle_channels_last
python test/test_proxy_tensor.py

cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10 @mlazos @soumith @voznesenskym @yanboliang @penguinwu @anijain2305 @EikanWang @Guobing-Chen @chunyuan-w @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @peterbell10 @desertfire

@pytorch-bot
Copy link

pytorch-bot bot commented Jan 3, 2023

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/91605

Note: Links to docs will display an error until the docs builds have been completed.

✅ No Failures

As of commit 8ceb796:
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@Xia-Weiwen Xia-Weiwen changed the title [Inductor] Reorder tensor inputs of FallbackKernel to expected strides [Meta] Register aten.pixel_shuffle.default for meta Jan 3, 2023
Copy link
Collaborator

@jgong5 jgong5 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. Add a UT to cover the fixed issue?

@Xia-Weiwen Xia-Weiwen added the intel This tag is for PR from Intel label Jan 4, 2023
@Xia-Weiwen
Copy link
Collaborator Author

LGTM. Add a UT to cover the fixed issue?

Hi @jgong5 I have added a UT. Please review again. Thanks!
It is noteworthy that looks like CUDA requires output always to be contiguous (channels-first) otherwise a UT fails. I don't know if it is true and why. I just made changes to let CI pass.

@Xia-Weiwen
Copy link
Collaborator Author

Hi @mingfeima Thanks for your approval. Does that mean this PR can be merged or can be set to ready for review from the Meta side?

@Xia-Weiwen Xia-Weiwen marked this pull request as ready for review January 5, 2023 02:06
@Xia-Weiwen
Copy link
Collaborator Author

Hi @jgong5 @mingfeima Thanks for your approval. Do we need approval from Meta side to land this PR? If so, from whom?

@Xia-Weiwen
Copy link
Collaborator Author

@pytorchbot merge

@pytorch-bot pytorch-bot bot added the ciflow/trunk Trigger trunk jobs on your pull request label Jan 5, 2023
@pytorchmergebot
Copy link
Collaborator

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging
Check the merge workflow status
here

@pytorchmergebot
Copy link
Collaborator

The merge job was canceled. If you believe this is a mistake,then you can re trigger it through pytorch-bot.

@Xia-Weiwen
Copy link
Collaborator Author

@pytorchbot merge

@pytorchmergebot
Copy link
Collaborator

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging
Check the merge workflow status
here

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ciflow/inductor ciflow/trunk Trigger trunk jobs on your pull request intel This tag is for PR from Intel Merged module: inductor open source

Projects

Status: Done

Development

Successfully merging this pull request may close these issues.

Pytorch 2.0 compiled model complains stride mismatch about pixel_shuffle

6 participants