workers flags not needed when reload is set to true.

Author: Jonathan Makunga

src/sagemaker/serve/app.py

@@ -74,7 +74,6 @@ def _create_server(self):
             log_level="info",
             loop="asyncio",
             reload=True,
-            workers=3,
             use_colors=True,
         )
 
@@ -91,11 +90,7 @@ def start_server(self):
 
     def stop_server(self):
         """Destroys the uvicorn server."""
-        if  self.is_running:
-            logger.info("Deleting server...")
-            # self._stop_event.set()
-            # self._thread.join()
-            logger.info("Server deleted.")
+        # TODO: Implement me.
 
     def _start_run_async_in_thread(self):
         loop = asyncio.new_event_loop()
@@ -104,6 +99,3 @@ def _start_run_async_in_thread(self):
 
     async def _serve(self):
         await self.server.serve()
-
-    def is_running(self):
-        return self._thread is not None and self._thread.is_alive()

src/sagemaker/serve/model_server/multi_model_server/server.py

@@ -2,9 +2,7 @@
 
 from __future__ import absolute_import
 
-import asyncio
 import json
-import threading
 
 import requests
 import logging
@@ -28,20 +26,9 @@
 class InProcessMultiModelServer:
     """In Process Mode Multi Model server instance"""
 
-    def __init__(self):
-        # from sagemaker.serve.app import InProcessServer
-        # self._in_process_server = InProcessServer
-        pass
-
-    # def run_async_in_thread(self):
-    #     loop = asyncio.new_event_loop()
-    #     asyncio.set_event_loop(loop)
-    #     loop.run_until_complete(self.main())
-
     def _start_serving(self):
         """Initializes the start of the server"""
         from sagemaker.serve.app import InProcessServer
-        # threading.Thread(target=self.run_async_in_thread, daemon=True).start()
         if hasattr(self, "inference_spec"):
             model_id = self.inference_spec.get_model()
         else:
@@ -58,7 +45,6 @@ def _invoke_multi_model_server_serving(self, request: bytes, content_type: str,
         """Placeholder docstring"""
         try:
             response = requests.post(
-                # "http://127.0.0.1:9007/generate",
                 f"http://{self.server.host}:{self.server.port}/generate",
                 data=request,
                 headers={"Content-Type": content_type, "Accept": accept},