changed to uvicorn setup for HF v15

main/api.py

@@ -29,8 +29,14 @@ class InferenceApi(LitAPI):
 
     async def _get_client(self):
         """Get or create HTTP client as needed"""
+        base_url = self.llm_config.get('base_url', 'http://localhost:8002')
+
+        # If it's a HF space URL, remove any port specification
+        if 'hf.space' in base_url:
+            base_url = base_url.split(':')[0]  # Take everything before any port number
+
         return httpx.AsyncClient(
-            base_url=self.llm_config.get('base_url', 'http://localhost:8001'),
+            base_url=base_url,
             timeout=float(self.llm_config.get('timeout', 60.0))
         )