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

[Model Runner V2] Remove unused warmup_for_prefill method (#36762)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk@inferact.ai>
This commit is contained in:
Woosuk Kwon
2026-03-11 08:02:44 -07:00
committed by GitHub
parent a9e532afe2
commit 8ccbcda5c0
1 changed files with 0 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
vllm/v1/worker/gpu/model_runner.py
View File

@@ -532,13 +532,6 @@ class GPUModelRunner(LoRAModelRunnerMixin):
) )
return cuda_graph_size return cuda_graph_size
def warmup_for_prefill(self) -> None:
# For FlashInfer, we would like to execute a dummy prefill run
# to trigger JIT compilation.
if all("FLASHINFER" in b.get_name() for b in self.attn_backends.values()):
self._dummy_run(self.max_num_tokens, skip_attn=False)
torch.accelerator.synchronize()
def finish_requests(self, scheduler_output: SchedulerOutput) -> None: def finish_requests(self, scheduler_output: SchedulerOutput) -> None:
finished_req_ids = scheduler_output.finished_req_ids finished_req_ids = scheduler_output.finished_req_ids
preempted_req_ids = scheduler_output.preempted_req_ids preempted_req_ids = scheduler_output.preempted_req_ids