Nvidia 发布 GR00T 2，这是一个可驱动机械臂、机械手和人形机器人的开放世界-动作模型。机器人通常依赖人工遥操作录制、将每帧画面与精确电机指令配对的数据，而普通人类视频缺少电机指令，此前模型难以利用。GR00T 2 通过预训练缓解这一问题，使用 402,730 小时人类头戴摄像头视频、10,184 小时手持夹爪演示，遥操作机器人数据仅 23,500 小时。
原文给出 GR00T 2 的预训练数据构成，读者可据此理解它为何能用低成本人类视频弥补机器人遥操作数据不足。
Nvidia released GR00T 2, an open world-action model that runs arms, robot hands and humanoids.
Robots normally learn from recordings that pair each camera frame with exact motor commands, and those only come from a person steering a real robot.
Ordinary human video is far cheaper and more plentiful, but it has no motor commands, so earlier models could barely use it.
GR00T 2 eases that by pretraining on 402,730 hours of human head-camera video plus 10,184 hours of handheld-gripper demos, but only 23,500 teleoperated robot hours.
来源：Rohan Paul · x.com