biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[V1] [KVConnector] Fix MultiprocExecutor worker output aggregation (#21048)

Browse Source 
Signed-off-by: David Ben-David <davidb@pliops.com>
Co-authored-by: David Ben-David <davidb@pliops.com>
This commit is contained in:
David Ben-David
2025-07-17 08:29:45 +03:00
committed by GitHub
parent 8a4e5c5f3c
commit 4fcef49ec4
2 changed files with 129 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/v1/executor/multiproc_executor.py
View File

@@ -273,10 +273,8 @@ class MultiprocExecutor(Executor):
output = outputs[self.output_rank]
# set the aggregated finished_sending / finished_recving
if finished_sending:
output.finished_sending = finished_sending
if finished_recving:
output.finished_recving = finished_recving
output.finished_sending = finished_sending if finished_sending else None
output.finished_recving = finished_recving if finished_recving else None
return output