return format change

app.py

@@ -73,7 +73,9 @@ def chat_fxn_caller(message, history, temperature=0.6, top_p=0.95, max_tokens=12
     if len(generated) > max_response_len:
         generated = generated[:max_response_len].rsplit(" ", 1)[0]
 
-    return generated
+    #return generated
+    return {"role": "assistant", "content": generated}
+
 
 demo = gr.ChatInterface(
     chat_fxn_caller,