demo_src_three_piece_assembly_task_D2_robot_IIWA_gripper_Robotiq85Gripper_additionalCams
收藏资源简介:
该数据集是一个由LeRobot工具创建的机器人学数据集,专注于机器人操作任务。它包含500个episodes,总计167,742帧数据,帧率为20fps。数据以parquet文件格式存储,视频以mp4格式存储。数据集的特征包括:动作(action,包含7个浮点数,代表dx、dy、dz、drx、dry、drz和夹爪控制)、观测状态(observation.state,包含2个浮点数)、末端执行器位置和姿态(observation.eef,包含7个浮点数,代表位置和四元数)、夹爪状态(observation.gripper,包含6个浮点数)、关节位置(observation.joint,包含7个浮点数),以及组合特征如observation.eef_gripper和observation.joint_gripper。此外,数据集提供来自前视图(frontview)和侧视图(sideview)的图像观测,每个图像分辨率为256x256,RGB格式,视频编码为av1。数据集适用于机器人学习任务,如装配、抓取或操作,但具体任务未在README中明确说明。数据集的许可证为Apache-2.0。
This dataset is a robotics dataset created by the LeRobot tool, focusing on robotic manipulation tasks. It contains 500 episodes, totaling 167,742 frames of data with a frame rate of 20fps. The data is stored in parquet file format, and videos are stored in mp4 format. The dataset features include: action (containing 7 floating-point numbers representing dx, dy, dz, drx, dry, drz, and gripper control), observation.state (containing 2 floating-point numbers), end-effector position and orientation (observation.eef, containing 7 floating-point numbers representing position and quaternion), gripper state (observation.gripper, containing 6 floating-point numbers), joint positions (observation.joint, containing 7 floating-point numbers), and combined features such as observation.eef_gripper and observation.joint_gripper. Additionally, the dataset provides image observations from frontview and sideview, with each image having a resolution of 256x256, RGB format, and video encoding in av1. The dataset is suitable for robotics learning tasks, such as assembly, grasping, or manipulation, but the specific task is not explicitly stated in the README. The dataset is licensed under Apache-2.0.
数据集概述
- 数据集名称:
OliverHausdoerfer/demo_src_three_piece_assembly_task_D2_robot_IIWA_gripper_Robotiq85Gripper_additionalCams - 许可证: Apache-2.0
- 任务类别: 机器人学 (robotics)
- 标签: LeRobot
- 创建工具: 使用 LeRobot 创建
数据集规模与结构
| 属性 | 数值 |
|---|---|
| 总片段数 (episodes) | 500 |
| 总帧数 (frames) | 167,742 |
| 总任务数 (tasks) | 1 |
| 分块大小 (chunk size) | 1,000 |
| 数据文件大小 | 100 MB |
| 视频文件大小 | 200 MB |
| 帧率 (fps) | 20 |
| 数据拆分 | 训练集: 0-500 片段 (全部用于训练) |
数据特征
数据集包含多种观测与动作特征,均为浮点型 (float32),具体如下:
| 特征名称 | 向量维度 | 说明 |
|---|---|---|
| action | 7 | 动作:dx, dy, dz, drx, dry, drz, gripper |
| observation.state | 2 | 状态:zero0, zero1 |
| observation.eef | 7 | 末端执行器位姿:位置 (x,y,z) + 四元数 (x,y,z,w) |
| observation.gripper | 6 | 夹爪关节位置:gripper_qpos_0 至 gripper_qpos_5 |
| observation.joint | 7 | 机器人关节位置:joint_pos_0 至 joint_pos_6 |
| observation.eef_gripper | 13 | 末端执行器位姿与夹爪位置的组合 |
| observation.joint_gripper | 13 | 关节位置与夹爪位置的组合 |
| observation.images.frontview | 视频 | 前视图,分辨率 256x256,RGB,AV1 编码,20fps |
| observation.images.sideview | 视频 | 侧视图,分辨率 256x256,RGB,AV1 编码,20fps |
| timestamp | 1 | 时间戳 |
| frame_index | 1 | 帧索引 |
| episode_index | 1 | 片段索引 |
| index | 1 | 全局索引 |
| task_index | 1 | 任务索引 |
数据存储格式
- 数据文件: Parquet 格式,路径为
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频文件: MP4 格式,路径为
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
其他信息
- 可视化: 可通过 Hugging Face Space 在线查看数据集。
- 引用: 暂无提供 BibTeX 信息。



