Open
Conversation
valarLip
reviewed
Jan 20, 2026
ChuanLi1101
reviewed
Jan 20, 2026
Collaborator
ChuanLi1101
left a comment
ChuanLi1101
left a comment
There was a problem hiding this comment.
Mostly good with minor suggestions.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
Aim was to create a proper solution that didn't just skip over the parameter for kv_scale or output_scale in LLFP4 or LLFP8,
and loaded each parameter properly.
Technical Details
Small changes to attention_mha to have k_scale and v_scale load as parameters not just NoneType, functions to remap the parameter names for the inputted tensors, and handling weight loading.
Test Plan
Trace looks normal, and lm_eval results match vllm lm_eval results with the same command used.
Test Result
Testing passed.
Submission Checklist