CLAMP: Contrastive Learning for 3D Multi-View Action-Conditioned Robotic Manipulation Pretraining figure
AlphaXiv 中文概览(可滚动查看)