Update app.py

app.py

@@ -189,17 +189,16 @@ async def respond_stream(message: str, chat_history: List[Dict[str, str]], max_t
     messages.append({"role": "user", "content": message})
     
     try:
-        stream = hf_client.text_generation(
-            "CohereForAI/c4ai-command-r-plus-08-2024",
+        stream = hf_client.chat_completion(
             messages,
-            max_new_tokens=max_tokens,
+            max_tokens=max_tokens,
             temperature=temperature,
             top_p=top_p,
             stream=True
         )
         
         full_response = ""
-        for response in stream:
+        async for response in stream:
             if response.token.special:
                 continue
             full_response += response.token.text
@@ -320,7 +319,7 @@ def create_ui():
                     msg = gr.Textbox(label="메시지")
             
                     max_tokens = gr.Slider(minimum=1, maximum=8000, value=4000, label="Max Tokens", visible=False)
-                    temperature = gr.Slider(minimum=0, maximum=1, value=0.7, label="Temperature", visible=False)
+                    temperature = gr.Slider(minimum=0, maximum=1, value=0.3, label="Temperature", visible=False)
                     top_p = gr.Slider(minimum=0, maximum=1, value=0.9, label="Top P", visible=False)
             
                     examples = [
@@ -331,20 +330,20 @@ def create_ui():
                         ["특허 출원에 활용할 기술 및 비즈니스모델 측면을 포함하여 특허 출원서 구성에 맞게 혁신적인 창의 발명 내용을 중심으로 4000토큰 이상 작성하라."],
                         ["계속 이어서 답변하라"],
                     ]
-        
+    
                     gr.Examples(examples, inputs=msg)
-        
+
                     def respond_wrapper(message, chat_history, max_tokens, temperature, top_p):
-                        bot_message = asyncio.run(respond_stream(message, chat_history, max_tokens, temperature, top_p).__anext__())
+                        generator = respond_stream(message, chat_history, max_tokens, temperature, top_p)
+                        bot_message = asyncio.run(generator.__anext__())
                         chat_history.append({"role": "user", "content": message})
                         chat_history.append({"role": "assistant", "content": bot_message})
                         return "", chat_history
-        
-                    msg.submit(respond_wrapper, [msg, chatbot, max_tokens, temperature, top_p], [msg, chatbot])
 
+                    msg.submit(respond_wrapper, [msg, chatbot, max_tokens, temperature, top_p], [msg, chatbot])
 
+            
 
-                
 
 
             space_id_state = gr.State()