ivs-demo-toilet-water-bottom-20260701

该数据集由LeRobot创建，专注于机器人学任务，具体针对ivs-demo-toilet-water-bottom-20260701场景。数据集包含13个episodes，总计4358帧数据，采样率为30fps。数据格式包括动作（action）、观测状态（observation.state）、末端执行器摄像头图像（observation.images.cam_ee）、点云数据（observation.point_cloud.cam_ee.xyz、mask、rgb）以及时间戳、帧索引等元数据。动作数据包含8个维度（如yaw.pos、pitch1.pos等），观测状态包含34个维度（如关节位置、速度、扭矩及末端执行器位姿）。图像数据分辨率为480x640，点云数据包含4096个点。数据集使用Apache 2.0许可证，适用于机器人控制、感知和模仿学习等研究。

This dataset was created by LeRobot and focuses on robotics tasks, specifically for the ivs-demo-toilet-water-bottom-20260701 scenario. It contains 13 episodes, totaling 4358 frames of data with a sampling rate of 30fps. The data format includes actions (action), observation states (observation.state), end-effector camera images (observation.images.cam_ee), point cloud data (observation.point_cloud.cam_ee.xyz, mask, rgb), and metadata such as timestamps and frame indices. Action data has 8 dimensions (e.g., yaw.pos, pitch1.pos), while observation state has 34 dimensions (e.g., joint positions, velocities, torques, and end-effector poses). Image data has a resolution of 480x640, and point cloud data contains 4096 points. The dataset uses the Apache 2.0 license and is suitable for research in robot control, perception, and imitation learning.