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 15, 11:00 UTC
[Misc] not show --model in vllm serve --help (#16691)
Signed-off-by: reidliu41 <reid201711@gmail.com> Co-authored-by: reidliu41 <reid201711@gmail.com>
Reid committed
May 11, 2025 at 16:47 UTC
009b3d53823053e3df50824fb9e087704b8631aa
1 file changed
+3
-1
vllm/engine/arg_utils.py
+3
-1
@@ -5,6 +5,7 @@ import argparse
5
import dataclasses
6
import json
7
import re
8
+import sys
9
import threading
10
import warnings
11
from dataclasses import MISSING, dataclass, fields, is_dataclass
@@ -440,7 +441,8 @@ class EngineArgs:
441
title="ModelConfig",
442
description=ModelConfig.__doc__,
443
)
443
- model_group.add_argument("--model", **model_kwargs["model"])
444
+ if 'serve' not in sys.argv[1:] and '--help' not in sys.argv[1:]:
445
+ model_group.add_argument("--model", **model_kwargs["model"])
446
model_group.add_argument("--task", **model_kwargs["task"])
447
model_group.add_argument("--tokenizer", **model_kwargs["tokenizer"])
448
model_group.add_argument("--tokenizer-mode",