update arg_utils.py

vllm/engine/arg_utils.py

@@ -811,14 +811,6 @@ class EngineArgs:
             default=EngineArgs.num_speculative_heads,
             help='The number of speculative heads to sample from '
                  'the draft model in speculative decoding.')
-        parser.add_argument('--model-loader-extra-config',
-                            type=nullable_str,
-                            default=EngineArgs.model_loader_extra_config,
-                            help='Extra config for model loader. '
-                            'This will be passed to the model loader '
-                            'corresponding to the chosen load_format. '
-                            'This should be a JSON string that will be '
-                            'parsed into a dictionary.')
         parser.add_argument(
             '--ignore-patterns',
             action="append",