Skip to content

Conversation

@gante
Copy link
Contributor

@gante gante commented Feb 18, 2025

What does this PR do?

See title.

Follows the same logic as in #36235: equivalence tests slow down development on the PT side.

In this particular case, this PR unblocks #36238: some models will start returning a Cache instance by default, which breaks the equivalence tests on those models.

cc @LysandreJik @ArthurZucker as we discussed on Slack; @Rocketknight1 fyi

@gante gante requested a review from ydshieh February 18, 2025 11:21
@HuggingFaceDocBuilderDev

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

Copy link
Collaborator

@ydshieh ydshieh left a comment

Choose a reason for hiding this comment

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

LGTM, but with a few nit questions. Thank you.

Are we going to remove pt_flax tests too?

@Rocketknight1
Copy link
Member

Happy to break the equivalence tests at this point!

@gante
Copy link
Contributor Author

gante commented Feb 19, 2025

@ydshieh

Are we going to remove pt_flax tests too?

Yes, right after this PR gets merged 😎 (to avoid making a massive PR, which tend to lower the quality of the review process)

@gante gante merged commit 0863eef into huggingface:main Feb 19, 2025
24 checks passed
@gante gante deleted the pt_tf_flax_equivalence branch February 19, 2025 11:55
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants