Repository navigation
[Fix][DLight] Decline GEMV reductions with only batch spatial axes - #20985
Merged
Merged
Conversation
tlopex
approved these changes
Nov 15, 2026
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Weighted centered reductions can look like GEMV when a broadcast operand uses every spatial axis, leaving only batch axes and triggering an assertion; decline those blocks so the reduction schedule handles them instead.
The regression verifies rejection and fallback, all 15 GEMV tests pass using the current pass wrapper with the retained compiler, and the patched compiler builds ECAPA for WebGPU with eight HF parity cases passing in Chrome.