fix(npu): support W4A8 expert weights in CAM MLP - #342
Merged
Merged
Conversation
Signed-off-by: bjf-frz <frz123db@gmail.com>
bjf-frz
requested review from
hsliuustc0106 and
jiangkuaixue123
as code owners
September 14, 2026 06:29
| ], | ||
| w2_scale=[experts.get_eplb_parameter("w2_weight_scale")], | ||
| ) | ||
| # ### PATCH START: W4A8 CAM expert weights |
Collaborator
There was a problem hiding this comment.
LGTM. This is AFD-owned code, so there is no need for the PATCH START/END markers here.
Contributor
Author
There was a problem hiding this comment.
LGTM. This is AFD-owned code, so there is no need for the PATCH START/END markers here.
Done, patch markers has been removed.
Signed-off-by: bjf-frz <frz123db@gmail.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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
CAM MLP needs the packed W4A8 weights, scales, and optional scale biases owned by the routed experts. Handle both static weights and dynamic EPLB weight lists, pass the per-channel quantization flag, and obtain the SwiGLU limit from the MLP wrapper.
Add 16 CPU contract tests covering dynamic EPLB, per-channel quantization, optional biases, and empty token batches. Update the existing quantization test mock for W4A8.
Validation:
The extracted change has not been deployed against the current upstream revision.