pickplace_vision
收藏资源简介:
该数据集是一个用于机器人任务的数据集,专门针对拾取放置视觉任务,使用LeRobot框架创建。数据集包含88个训练集,总计92752帧,数据以30帧每秒的速率采集。数据特征包括来自手腕和上方视角的观察图像(分辨率为120x160,3通道),观察状态(18维浮点向量),动作(9维浮点向量),以及时间戳、帧索引、集索引、索引和任务索引等元数据。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。机器人类型为Franka,任务数为1。
This dataset is designed for robotic tasks, specifically for pick-and-place visual tasks, and is created using the LeRobot framework. It contains 88 training sets with a total of 92,752 frames, collected at a rate of 30 frames per second. The data features include observation images from wrist and overhead viewpoints (resolution 120x160, 3 channels), observation states (18-dimensional floating-point vectors), actions (9-dimensional floating-point vectors), and metadata such as timestamps, frame indices, set indices, indices, and task indices. The data is stored in Parquet format, while videos are in MP4 format, with total data file size of 100MB and video file size of 200MB. The robot type is Franka, and the number of tasks is 1.
数据集概述:pickplace_vision
- 数据集名称:pickplace_vision
- 数据集地址:https://huggingface.co/datasets/jamongsteak/pickplace_vision
- 许可证:Apache-2.0
- 任务类别:机器人学(robotics)
- 相关标签:LeRobot
数据集描述
该数据集使用 LeRobot 创建,包含机器人操作任务的相关数据。
数据集结构
- 代码库版本:v3.0
- 机器人类型:Franka
- 总集数(Episodes):88
- 总帧数(Frames):92,752
- 总任务数(Tasks):1
- 块大小(Chunks Size):1,000
- 数据文件大小:100 MB
- 视频文件大小:200 MB
- 帧率(FPS):30
- 数据划分:训练集包含全部 88 个 episode
特征(Features)
| 特征名称 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
observation.images.wrist |
video | [120, 160, 3] | 腕部摄像头图像,分辨率120x160,AV1编码,30 FPS |
observation.images.over |
video | [120, 160, 3] | 俯视摄像头图像,分辨率120x160,AV1编码,30 FPS |
observation.state |
float32 | [18] | 机器人状态向量(18维) |
action |
float32 | [9] | 动作向量(9维) |
timestamp |
float32 | [1] | 时间戳 |
frame_index |
int64 | [1] | 帧索引 |
episode_index |
int64 | [1] | 集数索引 |
index |
int64 | [1] | 全局索引 |
task_index |
int64 | [1] | 任务索引 |
数据文件路径
- 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
数据结构可视化
数据集可以通过 LeRobot 可视化界面 进行预览。



