Skip to content

[Bugfix] Fix DeepSeek V2-Lite Accuracy drop #40673

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
robertgshaw2-redhat merged 4 commits into from
Apr 23, 2026
+9 −4
Merged

[Bugfix] Fix DeepSeek V2-Lite Accuracy drop #40673

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
33f7490
fix DeepSeek V2-Lite Accuracy
bnellnm Apr 23, 2026
0e2f7cd
better fix
bnellnm Apr 23, 2026
46295d9
update comment
bnellnm Apr 23, 2026
1deb55f
fix typo + DCO
bnellnm Apr 23, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 9 additions & 4 deletions vllm/model_executor/layers/fused_moe/runner/moe_runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -335,11 +335,16 @@ def _maybe_reduce_shared_expert_output(
"""All-reduce shared expert output when the combine kernel already
reduced fused output.

This is the "early" all-reduce path. When the combine kernel produces
already-reduced fused output, shared output must be reduced separately
to match.
* If the combine kernel does the reduction for fused_output, reduce
shared_output separately. O.w, reduce fused_output+shared_output later.
* If we have SP (TP=N, DP=M, EP), there is a separate AG step handled
in the model.
"""
if shared_output is not None and self._fused_output_is_reduced:
if (
shared_output is not None
and not self.moe_config.is_sequence_parallel
Comment thread
bnellnm marked this conversation as resolved.
and self._fused_output_is_reduced
):
shared_output = tensor_model_parallel_all_reduce(shared_output)
return shared_output

Expand Down
Loading