Update app.py

app.py

@@ -81,9 +81,10 @@ def respond(message, history, level, max_tokens, temperature, top_p):
     # Generate response
     response = ""
     try:
-        for msg in client.chat_completion(
-            messages, max_tokens=max_tokens, stream=True, temperature=temperature, top_p=top_p
-        ):
+        prompt = "\n".join([f"{m['role'].capitalize()}: {m['content']}" for m in messages])
+        for token in client.text_generation(prompt, max_new_tokens=max_tokens, stream=True, temperature=temperature, top_p=top_p):
+            response += token
+            yield response
             token = msg.choices[0].delta.content
             if token is not None:  # Handle None tokens
                 response += token