search

allenai / RL4LMs

A modular RL library to fine-tune language models to human preferences
https://rl4lms.apps.allenai.org/
Apache License 2.0
2.18k stars 191 forks source link

Mix-Precision training #29

Open lovodkin93 opened 1 year ago

lovodkin93 commented 1 year ago

Hey, Are there any plans to add support for mixed precision training? I did see in #12 a temporary solution was suggested, but it still throws multiple exceptions relating to mathematical operations between fp16 and fp32 values. Thanks! @rajcscw

rajcscw commented 1 year ago

Hey we are working on the support of hugging face's Accelerate. With that mixed precision training would be possible.

CathyKitten commented 9 months ago

May I ask if there is a complete code with changes that I can learn from？