tablewares_sort_ik5
收藏资源简介:
该数据集是一个用于机器人学任务的数据集,由LeRobot创建。它包含28个episodes,总计53703帧,用于单一任务。数据集记录了机器人(类型为openarm_follower)的动作和状态信息,包括8维动作向量(对应7个关节位置和夹爪位置)、8维状态观测(与动作相同),以及两个摄像头采集的图像数据:前视摄像头和腕部摄像头,均为分辨率720x1280的RGB视频(每秒30帧)。此外,数据集还包含时间戳、帧索引、episode索引、任务索引等元数据。数据以parquet格式存储,视频以mp4格式存储,总数据文件大小为100MB,视频文件大小为200MB。数据集适用于机器人控制、模仿学习或强化学习等研究。
This dataset is designed for robotics tasks and was created by LeRobot. It consists of 28 episodes, totaling 53,703 frames, for a single task. The dataset records the robots (type: openarm_follower) actions and state information, including an 8-dimensional action vector (corresponding to 7 joint positions and gripper position), an 8-dimensional state observation (same as the action), and image data captured by two cameras: a front-view camera and a wrist camera, both providing RGB video at a resolution of 720x1280 (30 frames per second). Additionally, the dataset includes metadata such as timestamps, frame indices, episode indices, and task indices. The data is stored in parquet format, with videos in mp4 format; the total data file size is 100MB, and the video file size is 200MB. The dataset is suitable for research in robot control, imitation learning, or reinforcement learning.
数据集概述
基本信息
- 数据集名称:tablewares_sort_ik5
- 许可证:Apache-2.0
- 任务类别:机器人学(robotics)
- 标签:LeRobot
数据集规模
- 总片段数:28
- 总帧数:53703
- 总任务数:1
- 帧率:30 FPS
- 块大小:1000
- 数据文件大小:100 MB
- 视频文件大小:200 MB
数据分割
- 训练集:片段 0 至 27(共28个片段)
机器人信息
- 机器人类型:openarm_follower
数据结构
数据集包含以下特征:
| 特征名称 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
| action | float32 | [8] | 7个关节位置 + 1个夹爪位置 |
| observation.state | float32 | [8] | 7个关节位置 + 1个夹爪位置 |
| observation.images.front | video | [720, 1280, 3] | 前方摄像头,分辨率720x1280,AV1编码,30 FPS |
| observation.images.wrist | video | [720, 1280, 3] | 腕部摄像头,分辨率720x1280,AV1编码,30 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] | 总体索引 |
| task_index | int64 | [1] | 任务索引 |
数据存储路径
- 数据文件:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频文件:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4



