Fix MPS crash on Apple Silicon with grouped-query attention#39
Merged
ssrajadh merged 1 commit intossrajadh:masterfrom Apr 12, 2026
Merged
Conversation
MPS SDPA crashes with "incompatible dimensions" when query and KV head counts differ (GQA), which affects both Qwen3-VL 8B and 2B models. Fall back to eager attention on MPS. Also disable a spurious transformers >=5.3 assertion that incorrectly reports video feature/token count mismatches on MPS. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Sulaiman Ghori <sulaiman.ghori@outlook.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
attn_implementation="eager"on MPS to avoid the buggy SDPA kerneltransformers >=5.3assertion that incorrectly reports video feature/token count mismatches on MPS (both counts are identical buttorch._checkfails)Test plan
--backend local --model qwen2bon Apple Silicon (MPS), ~5s/chunk🤖 Generated with Claude Code