Giters

huggingface / alignment-handbook

Robust recipes to align language models with human and AI preferences

Home Page:https://huggingface.co/HuggingFaceH4

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

DPO recipe saves a float32 model

tcapelle opened this issue · comments

commented

Hello,

I have been using the Zephry DPO recipe and the models I get are save in float32. I am using config_full and accelerate multi_gpu.yaml

I think the issue is that the config_full has not setup the model as bfloat

Should this be changed?