Update app.py

app.py

@@ -166,14 +166,26 @@ def threaded_chat_response(previous_chat: str, msg_prompt: str) -> dict:
     """
     generation_params = {"do_sample": True,"temperature": 0.7,"top_p": 0.95,"top_k": 40,"max_new_tokens": 512,"repetition_penalty": 1.1}
     pipe = pipeline("text-generation",model=model, tokenizer=tokenizer, **generation_params)
-    history_message= f''' incase you are asked about something you talked about the previous days or times or periods,  use this as reference to recall {previous_chat}. Do not point out everything just major keypoints. '''
-    try:
+    history_message= f''' incase you are asked about something you talked about the previous days or times or periods,  use this as reference to recall {previous_chat}. Do not point out everything just major keypoints. If nothing is mentioned about the past conversation ingore this. '''
+
+    if previous_chat:
+        
         prompt_template=f'''
         <|im_start|>system{system_message}<|im_end|>
         <|im_start|>user {msg_prompt}<|im_end|>
         <|im_start|>system{history_message}<|im_end|>
         <|im_start|>assistant
         '''
+    else:
+        prompt_template=f'''
+        <|im_start|>system{system_message}<|im_end|>
+        <|im_start|>user {msg_prompt}<|im_end|>
+        <|im_start|>assistant
+        '''
+
+    
+    try:
+        
         pipe_output = pipe(prompt_template)[0]['generated_text']
       
          # Separate user's prompt and assistant's response
@@ -301,6 +313,7 @@ async def chat_thread(request: Request, user_id: str):
 
         data = await request.json()
         msg_prompt = data.get('msg_prompt')
+        previous_chat=data.get('msg_history')
 
         if not msg_prompt:
             raise HTTPException(status_code=400, detail="Prompt not provided")