This is the repo of the paper: On Pre-Training for Visuo-Motor Control: Revisiting a Learning-from-Scratch Baseline The pretrained baselines contains MVP, PVR (Moco) and R3M. For R3M, we can just ...