fix: correct sign of periodic kernel expected gradient #936
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
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Outdated suggestions cannot be applied.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.
PR Type
Description
In current
main, the sign of the expected gradient for thePeriodicKernelis incorrect. However, this wasn't showing up in tests, since it was also evaluating to 0. On Jax 0.5.0, it was no longer evaluating to zero, and so started causing test failures. (See #930; this PR doesn't fix all the issues there.)We should probably investigate why the gradient was evaluating to 0, and why it isn't doing so in jax 0.5.0!
How Has This Been Tested?
The
grad_xandgrad_ytests pass with jax 0.5.0. Also, the definitions now match betweenTestPeriodicKernelandPeriodicKernel.Does this PR introduce a breaking change?
No.
Checklist before requesting a review