A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 13, 15:45 UTC
[Bug] Fix missing token_ids for reasoning parser models in chat completions #28246 (#28256)
[Bug] Fix missing token_ids for reasoning parser models in chat completions #28246 (#28256)
baonudesifeizhai committed
Nov 7, 2025 at 02:31 UTC
9da9208b202af05fe0f0a5623ee0d8ca969d7558
1 file changed
+3
vllm/entrypoints/openai/serving_chat.py
+3
@@ -1372,6 +1372,9 @@ class OpenAIServingChat(OpenAIServing):
1372
else "stop"
1373
),
1374
stop_reason=output.stop_reason,
1375
+ token_ids=(
1376
+ as_list(output.token_ids) if request.return_token_ids else None
1377
+ ),
1378
)
1379
choices.append(choice_data)
1380
continue