The simplest, fastest repository for training/finetuning medium-sized GPTs.

A repo for distributed training of language models with Reinforcement Learning via Human Feedback (RLHF)