README.md

robot-rlhf

Robot Learning through Human Feedback. Inspired by advancements in NLP, we train a robot policy via reinforcement learning using a reward function learned exclusively from human preferences.