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 14, 05:30 UTC
[Bugfix] remove duplicate tokens streamed in required tool choice streaming (#23312)
Signed-off-by: Jason Cheng <jasoncky96@gmail.com> Co-authored-by: Chauncey <chaunceyjiang@gmail.com>
Cheng Kuan Yong Jason committed
Sep 16, 2025 at 15:16 UTC
68dbde5dbb11b9250454d0c9f21a8b3da960b341
1 file changed
-3
vllm/entrypoints/openai/serving_chat.py
-3
@@ -828,9 +828,6 @@ class OpenAIServingChat(OpenAIServing):
828
history_tool_call_cnt += 1
829
tools_streamed[i] = True
830
831
- # update the previous values for the next iteration
832
- previous_texts[i] = current_text
833
-
831
# handle streaming deltas for tools with "auto" tool choice
832
# and reasoning parser
833
elif tool_choice_auto and self.reasoning_parser: