Fix: Perform sigmoid calculation in fp32 for aux loss stability #2765
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
yaox12
approved these changes
Jan 4, 2026
yaox12
added
the
Expert Review
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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 updates
compute_routing_scores_for_aux_lossto performsigmoidrouting score calculations infloat32, matching the existing behavior of thesoftmaxpath.Fixes #2741
Currently, the
softmaxrouting path explicitly casts tofloat32to avoid underflow/overflow in BF16/FP16. However, thesigmoidpath performs operations in the input dtype.While
sigmoidis bounded [0, 1], the subsequent normalization (scores / sum) involves accumulation that can suffer from precision loss in BF16, especially with a large number of experts. This change aligns both methods to use high-precision accumulation for auxiliary loss stability.