[Pytorch] add an option to disable TORCH_WARN and TORCH_WARN_ONCE log

[biubiuX]

Summary: Add an option to disable TORCH_WARN, some op could trigger spammy TOCH_WARN log which is not desired under certain scenario.

Test Plan:
Tested with
-pt.disable_warn = 1 and -pt.disable_warn = 0

verified TORCH_WARN and TORCH_WARN_ONCE are properly handled

tested with
-pt.strip_error_messages = 1, -pt.disable_warn = 0

verified strip error message is respected when warn is printed

Differential Revision: D40321550

[linux-foundation-easycla]

The committers listed above are authorized under a signed CLA.

• ✅ login: biubiuX / name: zxiong (b51436f)

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/87188

• 📄 Preview Python docs built from this PR
• 📄 Preview C++ docs built from this PR
• ❓ Need help or want to give feedback on the CI? Visit our office hours

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

✅ No Failures

As of commit f6ce0cb:
💚 Looks good so far! There are no failures yet. 💚

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

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D40321550

[kurtamohler]

Looks good to me, but there's a recent merge conflict

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D40321550

[biubiuX]

@kurtamohler mind taking another look ? thanks !

[kurtamohler]

Looks good, thanks!

Just to make sure you're aware, you can filter all warnings at the Python level like this:

import warnings
warnings.filterwarnings('ignore')

But this also filters out all warnings in your project, not just the PyTorch ones, so I can see why this DISABLE_WARN flag is useful.

There is a loose plan to add a warning filter to the C++ side of PyTorch (part of pytorch/rfcs#43). The details aren't decided yet, but we could potentially add a Python binding for it. If/when that exists, you could use that instead of having to build PyTorch yourself

[biubiuX]

Thanks for the tip. We're the consumer of the model and we don't directly work on the python code. We're working with our partner to fix/adjust warning, but it would be great if we also have log/warning config on c++ side.

[biubiuX]

not sure if the workflow failure is caused by my change, is there anything I can do here ?

[kurtamohler]

Yeah these failures don't seem to be related. Try rebasing just to make sure

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D40321550

[biubiuX]

Could anyone help me run the workflows ?

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D40321550

[kurtamohler]

The two failing jobs actually do seem to be caused by your changes. I think I may have steered you wrong earlier--I had only seen the Device or resource busy error in the logs and assumed it wasn't related. But there's another error that happens earlier in the logs:

C:\actions-runner\_work\pytorch\pytorch\c10/core/TensorImpl.h(1758): error C2275: 'c10::UserWarning': illegal use of this type as an expression
C:\actions-runner\_work\pytorch\pytorch\c10/core/TensorImpl.h(1758): note: see declaration of 'c10::UserWarning'
C:\actions-runner\_work\pytorch\pytorch\c10/core/TensorImpl.h(1758): error C2064: term does not evaluate to a function taking 0 arguments

I guess the macros are maybe acting slightly differently on windows builds? Not sure

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D40321550

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D40321550

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D40321550

[biubiuX]

@kurtamohler mind taking another look(also trigger the workflow) ? It indeed seems to be a subtle preprocessor bug in MVSC(gcc, clang works ok).

[biubiuX]

gentle ping @kurtamohler ; )

[kurtamohler]

@pytorchbot merge

[pytorchmergebot]

Merge failed

Reason: Approval needed from one of the following (Rule 'superuser'):
qxy11, chauhang, chenyang78, shunting314, soulitzer, ...

Details for Dev Infra team

Raised by workflow job

[kurtamohler]

Oh right, forgot we need @ezyang's approval before merging

[biubiuX]

@pytorchbot merge

[pytorchmergebot]

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