xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-04-26 09:27:07 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon 2025-08-25 08:55:12 -07:00
parent 65f93694be
commit b1d52734f7
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/v1/worker/gpu_model_runner.py
View File

@ -1968,7 +1968,7 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
num_nans_in_logits = {}
num_nans_for_index = logits.isnan().sum(dim=-1).cpu().numpy()
for i, req_id in input_batch.req_ids:
for i, req_id in enumerate(input_batch.req_ids):
num_nans_in_logits[req_id] = (int(num_nans_for_index[i])
if num_nans_for_index is not None
and i < logits.shape[0] else 0)