AtomVLA: Scalable Post-Training for Robotic Manipulation via Predictive Latent World Models
返回详情
VLA / Vision-Language-Action 每日论文卡