rollout_act_so101_2_040000_20260629_183151
收藏Hugging Face2026-06-29 更新2026-06-29 收录
下载链接:
https://huggingface.co/datasets/hattori317/rollout_act_so101_2_040000_20260629_183151
下载链接
链接失效反馈官方服务:
资源简介:
该数据集使用LeRobot创建,专注于机器人学任务。它包含一个episode,总帧数为8889,以30fps的帧率采集。数据特征包括动作(action)和观察(observation),其中动作和状态观察均涉及6个关节位置:shoulder_pan.pos、shoulder_lift.pos、elbow_flex.pos、wrist_flex.pos、wrist_roll.pos和gripper.pos,数据类型为float32。观察还包括OpenCV图像,分辨率为480x640,3通道,以视频格式存储(使用av1编解码器)。数据集结构包含时间戳、帧索引、episode索引等元数据,机器人类型为so_follower。数据以parquet文件格式组织,总数据文件大小为100MB,视频文件大小为200MB,适用于机器人控制和模仿学习等应用。
This dataset is created using LeRobot and focuses on robotics tasks. It contains one episode with a total of 8889 frames, collected at 30fps. The data features include actions and observations, where both actions and state observations involve six joint positions: shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, and gripper.pos, with data type float32. Observations also include OpenCV images with a resolution of 480x640, 3 channels, stored in video format (using the av1 codec). The dataset structure includes metadata such as timestamps, frame indices, and episode indices, with a robot type of so_follower. The data is organized in parquet file format, with a total data file size of 100MB and a video file size of 200MB, suitable for applications like robot control and imitation learning.
提供机构:
hattori317创建时间:
2026-06-29
原始信息汇总
数据集概述
- 名称:
hattori317/rollout_act_so101_2_040000_20260629_183151 - 许可证: Apache-2.0
- 任务类别: 机器人学 (robotics)
- 数据集格式: LeRobot 格式,数据文件为 Parquet,视频文件为 MP4
数据集结构
- 总片段数 (episodes): 1
- 总帧数: 8889
- 总任务数: 1
- 帧率: 30 FPS
- 机器人类型:
so_follower - 数据划分: 仅包含训练集 (train: 0:1)
特征 (Features)
| 特征名 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
action |
float32 | [6] | 6维动作指令(肩部、肘部、腕部、夹爪等关节位置) |
observation.state |
float32 | [6] | 观测到的6维关节状态(与动作空间一致) |
observation.images.OpenCV |
video | [480, 640, 3] | 来自OpenCV摄像头的视频帧,分辨率480x640,RGB三通道,编码为AV1 |
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 |
文件存储
- 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 - 块大小: 1000 帧
- 数据文件大小: 约 100 MB
- 视频文件大小: 约 200 MB
用途
该数据集适用于机器人模仿学习、策略训练(如 ACT 算法)等机器人学任务,提供了真实的机器人状态、动作指令以及第一视角视频观测数据。



