Robot-DIFT: Distilling Diffusion Features for Geometrically Consistent Visuomotor Control

返回详情 VLA / Vision-Language-Action 每日论文卡