add args

Author: AllentDan

lmdeploy/cli/serve.py

@@ -123,6 +123,10 @@ def add_parser_api_server():
                             type=str,
                             default='',
                             help='Qos policy config path')
+        parser.add_argument('--log-stats',
+                            type=bool,
+                            default=False,
+                            help='Whether log stats to prometheus')
         # common args
         ArgumentHelper.backend(parser)
         ArgumentHelper.log_level(parser)
@@ -294,7 +298,8 @@ def api_server(args):
                        log_level=args.log_level.upper(),
                        api_keys=args.api_keys,
                        ssl=args.ssl,
-                       qos_config_path=args.qos_config_path)
+                       qos_config_path=args.qos_config_path,
+                       log_stats=args.log_stats)
 
     @staticmethod
     def api_client(args):

lmdeploy/serve/async_engine.py

@@ -174,7 +174,7 @@ def __init__(self,
                                                 PytorchEngineConfig]] = None,
                  chat_template_config: Optional[ChatTemplateConfig] = None,
                  tp: int = 1,
-                 log_stats: bool = True,
+                 log_stats: bool = False,
                  **kwargs) -> None:
         logger.info(
             f'input backend={backend}, backend_config={backend_config}')

lmdeploy/serve/openai/api_server.py

@@ -993,6 +993,7 @@ def serve(model_path: str,
           api_keys: Optional[Union[List[str], str]] = None,
           ssl: bool = False,
           qos_config_path: str = '',
+          log_stats: bool = False,
           **kwargs):
     """An example to perform model inference through the command line
     interface.
@@ -1031,6 +1032,7 @@ def serve(model_path: str,
             a single api_key. Default to None, which means no api key applied.
         ssl (bool): Enable SSL. Requires OS Environment variables 'SSL_KEYFILE' and 'SSL_CERTFILE'.
         qos_config_path (str): qos policy config path
+        log_stats (bool): Whether log stats to prometheus.
     """ # noqa E501
     if os.getenv('TM_LOG_LEVEL') is None:
         os.environ['TM_LOG_LEVEL'] = log_level
@@ -1064,6 +1066,7 @@ def serve(model_path: str,
         backend_config=backend_config,
         chat_template_config=chat_template_config,
         tp=tp,
+        log_stats=log_stats,
         **kwargs)
 
     if qos_config_path: