I wonder why we need to preprocess LLama. I know why preprocess vicuna, because we will change the directory in config.json and use it, but I dont see anything related to LLama in the tutorial, so why?
I don't understand your problem... If you mean how to run our model, we have provided demo.ipynb. If you mean how to run LLaMA or Vicuna, I suggest reading the tutorials on HuggingFace.
I wonder why we need to preprocess LLama. I know why preprocess vicuna, because we will change the directory in config.json and use it, but I dont see anything related to LLama in the tutorial, so why?