[MRV2] Run pooling post processing on non-final PP ranks (#56666)

Signed-off-by: Taneem Ibrahim <taneem.ibrahim@gmail.com> Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>

Taneem Ibrahim committed Sep 14, 2026 at 18:11 UTC ba2ae9f23961ac67bc5c055da8c26fbd660989c6
1 file changed +2
vllm/v1/worker/gpu_worker.py
+2
@@ -1238,6 +1238,8 @@ class Worker(WorkerBase):
1238 )
1239 self._pp_send_work = handles[1:]
1240
1241 + if self.use_v2_model_runner and self.model_runner.is_pooling_model:
1242 + return self.model_runner.pool() # type: ignore
1243 return None
1244
1245 def take_draft_token_ids(self) -> DraftTokenIds | None: