Update app.py

app.py

@@ -33,7 +33,7 @@ def respond(message, history, system_message, max_tokens, temperature, top_p):
             prompt += f"\nUser: {user_msg} output: "
         if assistant_msg:
             prompt += f"\nAssistant: {assistant_msg}"
-    prompt += f"\nUser: {message} output:"
+    prompt += f"\n<user> input:{message} output:"
     
     # Tokenize the input prompt
     inputs = tokenizer(prompt, return_tensors="pt", truncation=True)