[None][perf] Explicitly set Kimi K3 router correction bias in FP32 - #17061
Closed
jiaganc wants to merge 1 commit into
Closed
[None][perf] Explicitly set Kimi K3 router correction bias in FP32#17061jiaganc wants to merge 1 commit into
jiaganc wants to merge 1 commit into
Conversation
Signed-off-by: Jiagan Cheng <jiaganc@nvidia.com>
jiaganc
changed the title
Collaborator
Author
|
Will implement in #17312. |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
@coderabbitai summary
Description
Make the Kimi K3 router correction-bias dtype invariant explicit by allocating
e_score_correction_biasas FP32. The checkpoint loader copies into thedestination parameter dtype, so the bias remains FP32 after loading.
Pass the stored parameter directly to the fused
noaux_tcrouting path insteadof requesting a per-call
.float()conversion. This preserves routing behaviorwhile ensuring the fused path cannot allocate an upcast tensor when model
construction uses a non-FP32 default dtype.
Test Coverage
pre-commit run --files tensorrt_llm/_torch/modules/kimi_k3_moe/kimi_k3_moe_gate.pypython3 -m py_compile tensorrt_llm/_torch/modules/kimi_k3_moe/kimi_k3_moe_gate.pyPR Checklist
Please review the following before submitting your PR:
PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.
PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.
Test cases are provided for new code paths (see test instructions)
If PR introduces API changes, an appropriate PR label is added - either
api-compatibleorapi-breaking. Forapi-breaking, includeBREAKINGin the PR title.Any new dependencies have been scanned for license and vulnerabilities
CODEOWNERS updated if ownership changes
Documentation updated as needed
Update tava architecture diagram if there is a significant design change in PR.
The reviewers assigned automatically/manually are appropriate for the PR.
Please check this after reviewing the above items as appropriate for this PR.
GitHub Bot Help
To see a list of available CI bot commands, please comment
/bot help.