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

[Model Runner V2] Minor simplification for finish_requests (#33048)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
This commit is contained in:
Woosuk Kwon
2026-01-25 18:35:02 -08:00
committed by GitHub
parent edf927bc9f
commit 2f0d3ba745
1 changed files with 6 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
vllm/v1/worker/gpu/model_runner.py
View File

@@ -416,13 +416,12 @@ class GPUModelRunner(LoRAModelRunnerMixin):
torch.cuda.synchronize()
def finish_requests(self, scheduler_output: SchedulerOutput) -> None:
if scheduler_output.preempted_req_ids is not None:
for req_id in scheduler_output.preempted_req_ids:
self.req_states.remove_request(req_id)
if self.supports_mm_inputs:
self.encoder_runner.remove_request(req_id)
self.prompt_logprobs_worker.remove_request(req_id)
for req_id in scheduler_output.finished_req_ids:
finished_req_ids = scheduler_output.finished_req_ids
if scheduler_output.preempted_req_ids:
finished_req_ids = finished_req_ids.union(
scheduler_output.preempted_req_ids
)
for req_id in finished_req_ids:
self.req_states.remove_request(req_id)
if self.supports_mm_inputs:
self.encoder_runner.remove_request(req_id)