Update app.py

app.py

@@ -3,7 +3,6 @@ import requests
 import os
 import json
 from collections import deque
-import asyncio
 import aiohttp
 
 TOKEN = os.getenv("HUGGINGFACE_API_TOKEN")
@@ -44,36 +43,27 @@ async def respond(
         "messages": messages,
         "stream": True
     }
-    
+
     async with aiohttp.ClientSession() as session:
         async with session.post("https://api-inference.huggingface.co/v1/chat/completions", headers=headers, json=payload) as response:
-            partial_words = ""
-            async for chunk in response.content:
-                if chunk:
-                    chunk_data = chunk.decode('utf-8')
-                    if chunk_data.startswith("data: "):
-                        chunk_data = chunk_data[6:]
-                    try:
+            try:
+                async for chunk in response.content:
+                    if chunk:
+                        chunk_data = chunk.decode('utf-8')
                         response_json = json.loads(chunk_data)
                         if "choices" in response_json:
                             delta = response_json["choices"][0].get("delta", {})
                             if "content" in delta:
                                 content = delta["content"]
-                                partial_words += content
-                                yield partial_words
-                    except json.JSONDecodeError:
-                        continue
+                                yield content
+            except json.JSONDecodeError:
+                continue
+            except StopAsyncIteration:
+                return  # 종료 신호를 적절히 처리
 
 theme = "Nymbo/Nymbo_Theme"
 
-css = """
-footer {
-    visibility: hidden;
-}
-"""
-
 demo = gr.ChatInterface(
-    css=css,
     fn=respond,
     theme=theme,
     additional_inputs=[
@@ -85,4 +75,4 @@ demo = gr.ChatInterface(
 )
 
 if __name__ == "__main__":
-    demo.queue().launch(max_threads=20)
+    demo.launch()