JEPA-VLA: Video Predictive Embedding is Needed for VLA Models

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