This PR allows 2min of response

Click on _Merge_ to add this feature

Most of the answers are cut

app.py

@@ -47,7 +47,7 @@ model = AutoModelForCausalLM.from_pretrained(
     device_map="auto",
     quantization_config=quantization_config)
 
-@spaces.GPU()
+@spaces.GPU(duration=120)
 def stream_chat(
     message: str, 
     history: list,