Perform shape checks for self.param AFTER unboxing #4079
Merged
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Outdated suggestions cannot be applied.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 pull request will unbox params before performing shape checks in flax/core/scope.py, rather than after.
This should not affect pjit / jax.jit users as unboxing generally doesn't touch traced array data (and wasn't designed for it). However, users that prefer per-device code and manual collectives will run into trouble when implementing certain patterns like FSDP through their own boxes (i.e., subclassing
nn.meta.AxisMetadata
to achieve a "delayed all-gather") because the shape check fixed here happens at the wrong time.Fixes # (issue) N/A
Checklist
checks if that's the case).
discussion (please add a
link).
documentation guidelines.
(No quality testing = no merge!)
All existing tests pass.