huggingface / huggingface-llama-recipes

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

How to run LLama8b/70b using FP8

vgoklani opened this issue · comments

Are the instructions available to converting to FP8?

I'd like to try converting both the 8B and 70B to FP8 and compare.

Thank you!

Hey @vgoklani

As far as I understand it, this notebook does just what you are asking for. 🤗