Update app.py

app.py

@@ -47,7 +47,6 @@ def convert_history(chat_history, max_input_length=1024):
         history_text = "<|assistant|>\n"
     return history_text
 
-@torch.inference_mode()
 @spaces.GPU
 def instruct(instruction, max_token_output=1024):
     input_text = instruction