skpro19/v2.4

该数据集是由LeRobot创建的机器人学数据集，用于机器人操作任务。它包含20个episodes，总计29940帧数据，帧率为30fps，机器人类型为lekiwi_client。数据集以parquet文件和视频文件形式存储，总数据文件大小为100MB，视频文件大小为200MB。特征包括动作（action）和观测状态（observation.state），其中动作和状态都包含10维浮点数，表示机器人右臂的关节位置（如肩部平移、肩部抬升、肘部弯曲、手腕弯曲、手腕滚动、夹爪位置）、速度（x、y、theta）以及升降轴高度。观测还包括多视角图像：头戴摄像头（head_top）、右手腕摄像头（wrist_right）和前视摄像头（front），每个图像分辨率为480x640，3通道RGB视频，编码为av1。此外，数据集还包含时间戳、帧索引、episode索引、索引和任务索引等元数据。所有数据仅用于训练，分割为train:0:20。

This dataset was created by LeRobot for robotics tasks, specifically for robot manipulation. It consists of 20 episodes, totaling 29,940 frames at 30 fps, with a robot type of lekiwi_client. The data is stored in parquet and video files, with a total data file size of 100 MB and video file size of 200 MB. Features include actions and observation states: both actions and states are 10-dimensional float arrays representing robot joint positions (e.g., arm right shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, gripper), velocities (x, y, theta), and lift axis height. Observations also include multi-view images from head-top, wrist-right, and front cameras, each with a resolution of 480x640, 3-channel RGB video encoded in av1. Additionally, the dataset contains metadata such as timestamp, frame index, episode index, index, and task index. All data is for training only, with splits defined as train:0:20.