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 11, 23:45 UTC
[typing] fix type (#29964)
Signed-off-by: Andy Xie <andy.xning@gmail.com>
Ning Xie committed
Dec 5, 2025 at 18:46 UTC
7ae13c66ba63a1e999d9a8939856bea3e6e152a0
2 files changed
+2
-2
vllm/reasoning/abs_reasoning_parsers.py
+1
-1
@@ -121,7 +121,7 @@ class ReasoningParser:
121
self,
122
original_tag: str | None,
123
tool_server: ToolServer | None,
124
- ) -> str:
124
+ ) -> str | None:
125
"""
126
Instance method that is implemented for preparing the structured tag
127
Otherwise, None is returned
vllm/reasoning/gptoss_reasoning_parser.py
+1
-1
@@ -145,7 +145,7 @@ class GptOssReasoningParser(ReasoningParser):
145
# This function prepares the structural tag to format reasoning output
146
def prepare_structured_tag(
147
self, original_tag: str | None, tool_server: ToolServer | None
148
- ) -> str:
148
+ ) -> str | None:
149
if original_tag is None:
150
if tool_server is None:
151
return json.dumps(no_func_reaonsing_tag)