merged-eraser-4tasks-frontonly
收藏资源简介:
该数据集是一个机器人控制数据集,使用LeRobot工具创建。它包含机器人动作和观察数据,具体包括6个关节(肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转和夹爪)的位置动作和状态观察,以及前视摄像头视频(分辨率720x1280,30fps,AV1编码)。数据集结构包括时间戳、帧索引、片段索引、任务索引等元数据。总共有384个片段、274,797帧和4个任务,数据以parquet格式存储,视频以mp4格式存储。机器人类型为so_follower,适用于机器人学习和控制任务。
This dataset is a robotic control dataset created using the LeRobot toolkit. It contains robotic action and observation data, specifically including positional actions and state observations for 6 joints (shoulder translation, shoulder elevation, elbow flexion, wrist flexion, wrist rotation, and gripper), as well as front-facing camera videos with a resolution of 720x1280, 30fps frame rate, and AV1 encoding. The dataset structure includes metadata such as timestamps, frame indices, episode indices, task indices, etc. In total, there are 384 episodes, 274,797 frames and 4 tasks. The structured data is stored in Parquet format, while the videos are stored in MP4 format. The robotic platform is so_follower, which is suitable for robotic learning and control tasks.
数据集概述
- 数据集名称:merged-eraser-4tasks-frontonly
- 许可证:Apache-2.0
- 任务类别:机器人学(robotics)
- 标签:LeRobot
数据集结构
- 总片段数:384
- 总帧数:274,797
- 总任务数:4
- 帧率:30 FPS
- 数据文件格式:Parquet
- 视频文件格式:MP4(编码为AV1,分辨率1280x720)
- 机器人类型:so_follower(软体跟随机器人)
特征说明
| 特征名称 | 数据类型 | 形状 | 描述 |
|---|---|---|---|
action |
float32 | [6] | 机器人动作指令,包括6个关节位置 |
observation.state |
float32 | [6] | 机器人状态观测,包括6个关节位置 |
observation.images.front |
视频 | [720, 1280, 3] | 前方摄像头采集的RGB图像 |
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 |
数据划分
- 训练集:全部384个片段(索引0-383)
文件存储
- 数据路径:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - 视频路径:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4 - 数据文件大小:约100 MB
- 视频文件大小:约200 MB
- 分块大小:1000帧/块
其他信息
- 创建工具:LeRobot(https://github.com/huggingface/lerobot)
- 可视化链接:https://huggingface.co/spaces/lerobot/visualize_dataset?path=namin72/merged-eraser-4tasks-frontonly



