EvoScene-VLA: Evolving Scene Beliefs Inside the Action Decoder for Chunked Robot Control

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