OpenRLHF icon indicating copy to clipboard operation
OpenRLHF copied to clipboard

Published 20 hours ago • verified publisher icon OpenLLMAI

PPO training configuration for train_ppo_llama.sh

Open MurrayTom opened this issue 2 months ago • 1 comments

Hello, I want to run train_ppo_llama.sh on 4 A100 80G. Do I need to reconfigure the GPU allocation of the 4 models?

MurrayTom avatar Apr 18 '24 03:04 MurrayTom

train_ppo_llama.sh shares gpus for all models, you just need to reconfigure the GPUs in train_ppo_llama_ray.sh

hijkzzz avatar Apr 18 '24 09:04 hijkzzz