Update app.py

app.py

@@ -25,8 +25,8 @@ class StopOnTokens(StoppingCriteria):
 
 
 system_role= 'system'
-user_role = 'assistant'
-assistant_role = "user"
+user_role = 'user'
+assistant_role = 'assistant'
 
 sft_start_token =  "<|im_start|>"
 sft_end_token = "<|im_end|>"
@@ -54,7 +54,7 @@ def predict(message, history):
     generate_kwargs = dict(
         model_inputs,
         streamer=streamer,
-        max_new_tokens=256,
+        max_new_tokens=512,
         do_sample=True,
         top_p= 0.75,
         top_k= 60,