so101-write-6
收藏Hugging Face2026-06-20 更新2026-06-20 收录
下载链接:
https://huggingface.co/datasets/yen-0/so101-write-6
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是一个机器人学相关数据集,使用LeRobot工具创建,专为单一任务设计。它包含10个剧集,总计4330帧数据,采用Apache 2.0许可证,适用于机器人控制和研究。数据结构包括动作、观察状态、图像(来自手腕和顶部视角的RGB视频)、时间戳、帧索引、剧集索引、索引和任务索引等特征。动作和观察状态均包含6个浮点值,分别对应机器人关节位置(如肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置)。图像数据以视频形式存储,手腕视角分辨率为480x640,顶部视角分辨率为480x848,帧率为30fps,使用AV1编解码器。数据以parquet文件格式组织,分块存储,总数据文件大小为100MB,视频文件大小为200MB。数据集仅包含训练集,适用于机器人模仿学习或强化学习任务。
This is a robotics-oriented dataset developed using the LeRobot toolkit, specifically tailored for a single dedicated task. It comprises 10 episodes with a total of 4330 frames of data, and is released under the Apache 2.0 license, suitable for robotics control and related research. The dataset's data structure includes multiple features: actions, observations, RGB videos captured from wrist and top-down perspectives, timestamps, frame indices, episode indices, indices, and task indices. Both actions and observations consist of 6 floating-point values, which respectively correspond to the robot's joint positions including shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist rotation, and gripper position. The image data is stored in video format: the videos from the wrist perspective have a resolution of 480×640, while those from the top-down perspective have a resolution of 480×848, with a frame rate of 30fps and encoded using the AV1 codec. The dataset is organized in Parquet file format with chunked storage; the total size of the Parquet data files is 100MB, and the total size of the video files is 200MB. The dataset only includes a training split, and is applicable to robot imitation learning or reinforcement learning tasks.
提供机构:
yen-0创建时间:
2026-06-20
原始信息汇总
数据集概述:so101-write-6
- 许可证: Apache-2.0
- 任务类别: 机器人学 (Robotics)
- 标签: LeRobot
- 创建工具: 该数据集使用 LeRobot 创建。
数据集结构
- 机器人类型: so_follower
- 总片段数: 10
- 总帧数: 4330
- 总任务数: 1
- 数据块大小: 1000
- 数据文件大小: 100 MB
- 视频文件大小: 200 MB
- FPS: 30
- 数据集划分: 训练集 (train) 包含所有 10 个片段 (索引 0 到 9)
数据特征
| 特征名称 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
action |
float32 | [6] | 动作数据,包含6个关节位置 (shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper) |
observation.state |
float32 | [6] | 机器人状态观测,包含6个关节位置 (与 action 相同) |
observation.images.wrist |
video | [480, 640, 3] | 腕部摄像头视频,分辨率 480x640,AV1 编码,30 FPS |
observation.images.top |
video | [480, 848, 3] | 顶部摄像头视频,分辨率 480x848,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] | 全局索引 |
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



