diff --git a/fastdeploy/engine/common_engine.py b/fastdeploy/engine/common_engine.py
index 62442d8cb..41c12a2d6 100644
--- a/fastdeploy/engine/common_engine.py
+++ b/fastdeploy/engine/common_engine.py
@@ -589,7 +589,7 @@ class EngineSevice:
                 else:
                     err, data = self.zmq_server.receive_pyobj_once(block)
                 if err is not None:
-                    llm_logger.error("Engine stops inserting zmq task into scheduler, err:{err}")
+                    llm_logger.error(f"Engine stops inserting zmq task into scheduler, err:{err}")
                     break
 
                 request, insert_task = None, []
diff --git a/fastdeploy/entrypoints/openai/api_server.py b/fastdeploy/entrypoints/openai/api_server.py
index 4764ad9c7..16d6e735b 100644
--- a/fastdeploy/entrypoints/openai/api_server.py
+++ b/fastdeploy/entrypoints/openai/api_server.py
@@ -60,6 +60,7 @@ from fastdeploy.utils import (
     StatefulSemaphore,
     api_server_logger,
     console_logger,
+    is_package_installed,
     is_port_available,
     retrive_model_from_server,
 )
@@ -67,7 +68,7 @@ from fastdeploy.utils import (
 parser = FlexibleArgumentParser()
 parser.add_argument("--port", default=8000, type=int, help="port to the http server")
 parser.add_argument("--host", default="0.0.0.0", type=str, help="host to the http server")
-parser.add_argument("--workers", default=1, type=int, help="number of workers")
+parser.add_argument("--workers", default=None, type=int, help="number of workers")
 parser.add_argument("--metrics-port", default=8001, type=int, help="port for metrics server")
 parser.add_argument("--controller-port", default=-1, type=int, help="port for controller server")
 parser.add_argument(
@@ -82,6 +83,15 @@ parser.add_argument(
 )
 parser = EngineArgs.add_cli_args(parser)
 args = parser.parse_args()
+
+if args.workers is None:
+    # In GPU, the workers of uvicorn will be set according to the parameter `max-num-seqs`
+    if is_package_installed("paddlepaddle-gpu"):
+        args.workers = max(min(int(args.max_num_seqs // 32), 8), 1)
+    else:
+        args.workers = 1
+console_logger.info(f"Number of api-server workers: {args.workers}.")
+
 args.model = retrive_model_from_server(args.model, args.revision)
 chat_template = load_chat_template(args.chat_template, args.model)
 if args.tool_parser_plugin:
diff --git a/fastdeploy/utils.py b/fastdeploy/utils.py
index ecefd87af..2d939fa41 100644
--- a/fastdeploy/utils.py
+++ b/fastdeploy/utils.py
@@ -27,6 +27,7 @@ import sys
 import tarfile
 import time
 from datetime import datetime
+from importlib.metadata import PackageNotFoundError, distribution
 from logging.handlers import BaseRotatingHandler
 from pathlib import Path
 from typing import Literal, TypeVar, Union
@@ -668,6 +669,14 @@ def import_from_path(module_name: str, file_path: Union[str, os.PathLike]):
     return module
 
 
+def is_package_installed(package_name):
+    try:
+        distribution(package_name)
+        return True
+    except PackageNotFoundError:
+        return False
+
+
 def version():
     """
     Prints the contents of the version.txt file located in the parent directory of this script.