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 5, 18:30 UTC
[Misc] Move --grpc CLI argument into make_arg_parser (#38570)
Signed-off-by: Chang Su <chang.s.su@oracle.com>
Chang Su committed
Mar 31, 2026 at 03:24 UTC
acf7292bf239b9eaec9fc9d4383b0877a3f066e1
2 files changed
+7
-7
vllm/entrypoints/cli/serve.py
-7
@@ -136,13 +136,6 @@ class ServeSubcommand(CLISubcommand):
136
)
137
138
serve_parser = make_arg_parser(serve_parser)
139
- serve_parser.add_argument(
140
- "--grpc",
141
- action="store_true",
142
- default=False,
143
- help="Launch a gRPC server instead of the HTTP OpenAI-compatible "
144
- "server. Requires: pip install vllm[grpc].",
145
- )
139
serve_parser.epilog = VLLM_SUBCMD_PARSER_EPILOG.format(subcmd=self.name)
140
return serve_parser
141
vllm/entrypoints/openai/cli_args.py
+7
@@ -345,6 +345,13 @@ def make_arg_parser(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
345
"Must be a YAML with the following options: "
346
"https://docs.vllm.ai/en/latest/configuration/serve_args.html",
347
)
348
+ parser.add_argument(
349
+ "--grpc",
350
+ action="store_true",
351
+ default=False,
352
+ help="Launch a gRPC server instead of the HTTP OpenAI-compatible "
353
+ "server. Requires: pip install vllm[grpc].",
354
+ )
355
parser = FrontendArgs.add_cli_args(parser)
356
parser = AsyncEngineArgs.add_cli_args(parser)
357