Update app.py

app.py

@@ -18,9 +18,8 @@ class StopOnTokens(StoppingCriteria):
         for stop_id in stop_ids:
             if input_ids[0][-1] == stop_id:
                 return True
-        return
- False
-
+        return False
+model.to('cuda')
 @spaces.GPU(duration=195)
 def predict(message, history, temperature, max_tokens, top_p, top_k):
     history_transformer_format = history + [[message, ""]]