addressing the issue #34611 to make FlaxDinov2 compatible with any batch size #35138
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
MHRDYN7
changed the title
Contributor
Author
|
@NielsRogge, since you've been working on Dinov2 registers recently, can I please request a quick review on this one as well? |
Contributor
Author
|
@gante I'm not really sure who is currently working on flax in the hf team, could you please mention the right person? Need a quick review to merge this |
gante
approved these changes
Feb 25, 2025
Contributor
gante
left a comment
gante
left a comment
There was a problem hiding this comment.
The changes look reasonable to me, thank you for having a look at the issue 🤗 And thank you for modifying the tests to prevent regressions 🙌
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR solves the batch size issue with FlaxDinov2 model #34611. The commonly used tests across the transformers library could not detect the error in the "interpolate_position_encoding" method, which only worked for single images. The issue has been fixed using jnp.tile for it's simplicity, although jnp.repeat could also be used.
The slow tests have also been modified to pass a batch of images instead of just one.
@amyeroberts could you please review the changes.
P.S. this pr does exactly what #34620 aimed to do, however, with a few more improvements.