xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-01-07 23:00:54 +08:00
Code Issues Packages Projects Releases Wiki Activity

[ux] Switch a warning to debug about a pytorch fallback (#23750)

Browse Source 
Signed-off-by: Russell Bryant <rbryant@redhat.com>
Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
Russell Bryant 2025-09-25 10:38:16 -04:00 committed by yewentao256
parent 2469b8291b
commit 054c8b526f
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
vllm/v1/sample/ops/topk_topp_sampler.py
View File

@ -109,9 +109,9 @@ class TopKTopPSampler(nn.Module):
# CPU-GPU synchronization while `flashinfer_sample` does.
if (k is None and p is None) or generators:
if generators:
logger.warning_once("FlashInfer 0.2.3+ does not support "
"per-request generators. Falling back to "
"PyTorch-native implementation.")
logger.debug_once("FlashInfer 0.2.3+ does not support "
"per-request generators. Falling back to "
"PyTorch-native implementation.")
return self.forward_native(logits, generators, k, p)
assert self.logprobs_mode not in (
"processed_logits", "processed_logprobs"