Update sparse.grad() to support re-packing gradients into PyTrees #19760
Conversation
This file contains 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
Blair-Johnson
changed the title
Blair-Johnson
changed the title
|
@jakevdp This PR is ready for review if you get the chance. |
jakevdp
requested changes
Apr 16, 2024
jakevdp
approved these changes
Apr 19, 2024
google-ml-butler
bot
added
kokoro:force-run
pull ready
jakevdp
requested changes
Apr 19, 2024
jakevdp
requested changes
Apr 19, 2024
|
@jakevdp Finally getting back around to this, apologies for the glacial pace. I adopted your suggestions with some minor fixes. |
|
Hi - we're seeing a couple test failures here that have been fixed on the main branch. Can you rebase against the most recent main branch commit so we can try again? |
… pytrees & test cases
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Fixes #16582
This PR modifies the postprocessing step of
sparse.grad()to reconstruct the input PyTrees that were indexed for autodiff. Previously, only the gradient corresponding to the first element of a PyTree would be returned.This PR includes several test cases to verify the behavior of
sparse.grad()matches that ofjax.grad()when gradients are taken with respect to pytrees.