Reducing the duration under allowed maximum

app.py

@@ -52,7 +52,7 @@ class StopOnSequence(StoppingCriteria):
         )
 
 
-@spaces.GPU(duration=512)
+@spaces.GPU(duration=42)
 def spaces_model_predict(message: str, history: list[tuple[str, str]]):
     history_transformer_format = history + [[message, ""]]
     stop = StopOnSequence("<|human|>", tokenizer)