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, 22:35 UTC
[Bugfix] we should use metavar is not choices (#21902)
Signed-off-by: rongfu.leng <rongfu.leng@daocloud.io>
rongfu.leng committed
Jul 30, 2025 at 21:51 UTC
da3e0bd6e53f12bb18d518940e8150ba023956aa
1 file changed
+3
-1
vllm/entrypoints/openai/cli_args.py
+3
-1
@@ -194,7 +194,9 @@ schema. Example: `[{"type": "text", "text": "Hello world!"}]`"""
194
195
# Special case: Tool call parser shows built-in options.
196
valid_tool_parsers = list(ToolParserManager.tool_parsers.keys())
197
- frontend_kwargs["tool_call_parser"]["choices"] = valid_tool_parsers
197
+ parsers_str = ",".join(valid_tool_parsers)
198
+ frontend_kwargs["tool_call_parser"]["metavar"] = (
199
+ f"{{{parsers_str}}} or name registered in --tool-parser-plugin")
200
201
frontend_group = parser.add_argument_group(
202
title="Frontend",