Robometer: Scaling General-Purpose Robotic Reward Models via Trajectory Comparisons
Published:
Robometer is a general-purpose, video-language-input, dense reward model trained on RBM-1M, a large-scale dataset of over 1M trajectories spanning 21 robot embodiments. It significantly improves robot learning across online RL, offline RL, model-based RL, failure detection, and data retrieval for imitation learning. Link redirects to project website.
