demo_src_stack_three_task_D0_robot_Panda_gripper_PandaGripper_additionalCams
收藏Hugging Face2026-06-22 更新2026-06-22 收录
下载链接:
https://huggingface.co/datasets/OliverHausdoerfer/demo_src_stack_three_task_D0_robot_Panda_gripper_PandaGripper_additionalCams
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是一个用于机器人学研究的跨本体数据集,由LeRobot工具创建。它包含500个训练episodes,总计128,334帧数据,帧率为20 fps。数据涉及Panda机器人抓手的操作任务,包括多种观测特征:动作(7维浮点数组,表示dx、dy、dz、drx、dry、drz和gripper控制)、观测状态(2维浮点数组)、末端执行器位置和四元数(7维)、抓手关节位置(6维)、关节位置(7维),以及组合特征如末端执行器抓手和关节抓手。此外,数据集提供前视图和侧视图的图像观测,每帧图像分辨率为256x256 RGB,以视频格式存储。数据以parquet文件组织,视频以mp4格式存储,总数据大小约100 MB,视频大小约200 MB。数据集适用于机器人控制、模仿学习等任务。
This dataset is a cross-embodiment dataset for robotics research, created by the LeRobot tool. It contains 500 training episodes, totaling 128,334 frames of data with a frame rate of 20 fps. The data involves manipulation tasks with a Panda robot gripper, including various observation features: actions (7-dimensional float arrays representing dx, dy, dz, drx, dry, drz, and gripper control), observation states (2-dimensional float arrays), end-effector position and quaternion (7-dimensional), gripper joint positions (6-dimensional), joint positions (7-dimensional), and combined features such as end-effector gripper and joint gripper. Additionally, the dataset provides image observations from front and side views, with each frame having a resolution of 256x256 RGB and stored in video format. The data is organized in parquet files, videos are stored in mp4 format, with a total data size of approximately 100 MB and video size of about 200 MB. The dataset is suitable for tasks such as robot control and imitation learning.
提供机构:
OliverHausdoerfer创建时间:
2026-06-22
原始信息汇总
数据集概述
- 数据集名称:
demo_src_stack_three_task_D0_robot_Panda_gripper_PandaGripper_additionalCams - 许可证: Apache-2.0
- 任务类别: 机器人学 (Robotics)
- 创建工具: 基于 LeRobot 创建
数据集结构
- 代码库版本: v3.0
- 机器人类型: 跨实体 (cross_embodiment)
- 总片段数 (episodes): 500
- 总帧数 (frames): 128,334
- 总任务数 (tasks): 1
- 块大小 (chunks): 1000
- 数据文件大小: 100 MB
- 视频文件大小: 200 MB
- 帧率 (FPS): 20
- 数据集划分:
- 训练集: 片段 0 到 499(共 500 个片段)
数据特征
数据集包含以下特征列,每个特征列的数据类型、形状及命名规范如下:
| 特征名称 | 数据类型 | 形状 | 字段说明 |
|---|---|---|---|
action |
float32 | [7] | 动作指令:位移 (dx, dy, dz)、旋转 (drx, dry, drz)、夹爪 (gripper) |
observation.state |
float32 | [2] | 状态观测:zero0, zero1 |
observation.eef |
float32 | [7] | 末端执行器位姿:位置 (x, y, z) 和四元数 (quat_x, quat_y, quat_z, quat_w) |
observation.gripper |
float32 | [6] | 夹爪关节位置:gripper_qpos_0 至 gripper_qpos_5 |
observation.joint |
float32 | [7] | 关节位置:joint_pos_0 至 joint_pos_6 |
observation.eef_gripper |
float32 | [13] | 末端执行器位姿与夹爪关节位置的组合 (7 + 6) |
observation.joint_gripper |
float32 | [13] | 关节位置与夹爪关节位置的组合 (7 + 6) |
observation.images.frontview |
视频 | [256, 256, 3] | 前视图图像:RGB,分辨率 256x256,AV1 编码,20 FPS |
observation.images.sideview |
视频 | [256, 256, 3] | 侧视图图像:RGB,分辨率 256x256,AV1 编码,20 FPS |
timestamp |
float32 | [1] | 时间戳 |
frame_index |
int64 | [1] | 帧索引 |
episode_index |
int64 | [1] | 片段索引 |
index |
int64 | [1] | 全局索引 |
task_index |
int64 | [1] | 任务索引 |
数据路径
- 数据文件路径:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频文件路径:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
可视化
可通过以下按钮在线可视化该数据集: 可视化该数据集



