Optimize MoEs for decoding using batched_mm#43126
Merged
IlyasMoutawwakil merged 15 commits intomainfrom Jan 7, 2026
Merged
Conversation
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
IlyasMoutawwakil
commented
Jan 6, 2026
IlyasMoutawwakil
commented
Jan 6, 2026
IlyasMoutawwakil
commented
Jan 6, 2026
IlyasMoutawwakil
commented
Jan 6, 2026
ArthurZucker
approved these changes
Jan 6, 2026
Collaborator
ArthurZucker
left a comment
There was a problem hiding this comment.
SGTM! good that its a warning, it could be info as well !
cc @remi-or for CB we need to check perfs as well + using kernels
Contributor
|
[For maintainers] Suggested jobs to run (before merge) run-slow: deepseek_v2, ernie4_5_moe, flex_olmo, qwen2_moe |
Merged
2 tasks
SangbumChoi
pushed a commit
to SangbumChoi/transformers
that referenced
this pull request
Jan 23, 2026
* optimize model for decoding * only optimize when grouped_mm * fixes * fix training compile failures * no need to skip * style * fix * Apply suggestion from @IlyasMoutawwakil * Apply suggestion from @IlyasMoutawwakil * info once
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
3 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
What does this PR do?
batched_mm is more performant on "smaller inputs", especially when used with torch.compile.
smaller inputs: up to 2^6-2^9 tokens depending on the num experts and num topk
Before submitting
Pull Request section?
to it if that's the case.
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.