Goal-VLA: Image-Generative VLMs as Object-Centric World Models Empowering Zero-shot Robot Manipulation

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