Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Fix Test] Fix failed generation test by pining pytorch nightlies #2362

Merged
merged 3 commits into from
Feb 8, 2025

Conversation

acisseJZhong
Copy link
Contributor

@acisseJZhong acisseJZhong commented Feb 7, 2025

Context

What is the purpose of this PR? Is it to

  • add a new feature
  • fix a bug
  • update tests and/or documentation
  • other (please add here)

Please link to any issues this PR addresses.

Some of the generation tests are failing with latest pytorch nightlies. See link. We(thanks to @ebsmothers ) triaged the issue down to pytorch nightlies 2/2. To fix CI for now, we pinged pytorch nightlies to the 2/1 version.

Need to look into what exactly in 2/2 causes the failure.

Changelog

What are the changes made in this PR?
*

Test plan

Please make sure to do each of the following if applicable to your PR. If you're unsure about any one of these just ask and we will happily help. We also have a contributing page for some guidance on contributing.

CI for nightly green:
image

  • run pre-commit hooks and linters (make sure you've first installed via pre-commit install)
  • add unit tests for any new functionality
  • update docstrings for any new or updated methods or classes
  • run unit tests via pytest tests
  • run recipe tests via pytest tests -m integration_test
  • manually run any new or modified recipes with sufficient proof of correctness
  • include relevant commands and any other artifacts in this summary (pastes of loss curves, eval results, etc.)

UX

If your function changed a public API, please add a dummy example of what the user experience will look like when calling it.
Here is a docstring example
and a tutorial example

  • I did not change any public API
  • I have added an example to docs or docstrings

Copy link

pytorch-bot bot commented Feb 7, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/torchtune/2362

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

⏳ No Failures, 10 Pending

As of commit 085515c with merge base 9b38360 (image):
💚 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 facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Feb 7, 2025
@@ -46,7 +46,7 @@ jobs:
run: python -m pip install --upgrade pip
- name: Install torch nightly
if: ${{ matrix.torch-version == 'nightly' }}
run: python -m pip install --pre torch torchvision torchao --index-url https://download.pytorch.org/whl/nightly/cu126
run: python -m pip install --pre torch==2.7.0.dev20250201 torchvision==2.7.0.dev20250201 torchao==2.7.0.dev20250201 --index-url https://download.pytorch.org/whl/nightly/cu126
Copy link
Contributor

Choose a reason for hiding this comment

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

I think ao and torchvision versions are not right here

Copy link
Contributor Author

Choose a reason for hiding this comment

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

oops got it!

Copy link
Contributor

@ebsmothers ebsmothers left a comment

Choose a reason for hiding this comment

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

Thanks for doing this! The one tricky thing is testing.. we want to make sure that this will unbreak our CI, but the CI against PyTorch nightlies doesn't run until push. So can you do the following?

  1. Comment out L33-34 of gpu_test.yaml, push changes, let CI run
  2. Confirm that CI is green, then revert to the original version (i.e. uncomment L33-34). Then we can push

(1) will allow the CI to also run on nightlies on this PR so we can get signal before landing

Copy link
Contributor

@ebsmothers ebsmothers left a comment

Choose a reason for hiding this comment

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

Thank you for the fix! Looking forward to having green CI again

@acisseJZhong acisseJZhong merged commit 8c9235e into pytorch:main Feb 8, 2025
17 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants