โ
Code
I'm using mistral-7b to understand LLMs' procedure.
Does anyone have an idea to improve this process?
do not recommend changing the number of tokens -> 1. :)