salmonHan/record-test_20260528_010919
收藏资源简介:
该数据集使用LeRobot创建,专注于机器人学领域,具体针对追随者类型的机器人(robot_type: so_follower)。数据集包含1个任务(total_tasks: 1)和1个完整情节(total_episodes: 1),总计1797帧数据(total_frames: 1797),帧率为30fps。数据以Parquet文件格式存储,总数据文件大小为100MB,视频文件大小为200MB。数据集特征包括:动作(action)和观测状态(observation.state),均为6维浮点数组,对应机器人关节位置(如肩部平移、肩部抬升、肘部弯曲等);观测图像(observation.images.front)为视频格式,分辨率720x1280,3个通道;以及时间戳、帧索引、情节索引等元数据。数据集结构支持训练分割(splits: train),适用于机器人控制和学习任务。
This dataset was created using LeRobot and focuses on the robotics domain, specifically for a follower type robot (robot_type: so_follower). It contains 1 task (total_tasks: 1) and 1 complete episode (total_episodes: 1), with a total of 1797 frames (total_frames: 1797) at a frame rate of 30fps. The data is stored in Parquet format, with a total data file size of 100MB and video file size of 200MB. Dataset features include: action and observation.state, both as 6-dimensional float arrays corresponding to robot joint positions (e.g., shoulder pan, shoulder lift, elbow flex); observation.images.front as video format with 720x1280 resolution and 3 channels; and metadata such as timestamp, frame index, and episode index. The dataset structure supports a training split (splits: train) and is suitable for robot control and learning tasks.




