SingleTask
收藏资源简介:
该数据集是一个机器人学相关数据集,专门针对omy_f3m机器人类型,使用Physical AI Tools和LeRobot工具创建。数据集包含25个episodes,总计19149帧和75个视频,帧率为30fps。数据以parquet格式存储,仅包括训练集(涵盖所有25个episodes)。数据集特征丰富,包括时间戳、帧索引、episode索引、任务索引,以及观察数据(来自三个摄像头的图像:cam_wrist(分辨率480x848x3)、cam_top(分辨率480x640x3)、cam_side(分辨率480x640x3),以及7个关节状态)和动作数据(7个关节动作)。这些数据适用于机器人控制、行为学习等研究场景。
This is a robotics dataset specifically designed for the omy_f3m robot type, created using Physical AI Tools and LeRobot. The dataset contains 25 episodes, totaling 19149 frames and 75 videos, with a frame rate of 30 fps. It is stored in Parquet format and only includes the training split, which covers all 25 episodes. The dataset has rich features, including timestamps, frame index, episode index, task index, as well as observation data (images from three cameras: cam_wrist with a resolution of 480×848×3, cam_top with a resolution of 480×640×3, cam_side with a resolution of 480×640×3, plus 7 joint states) and action data (7 joint actions). This dataset is suitable for research scenarios such as robot control and behavior learning.
数据集概述
数据集名称: SingleTask
数据集地址: https://huggingface.co/datasets/Syntax-Terror-BV/SingleTask
许可证: Apache-2.0
任务类别: 机器人学 (Robotics)
标签: robotis、LeRobot、omy_f3m
创建工具: 本数据集使用 Physical AI Tools 和 LeRobot 创建。
数据集结构
| 属性 | 值 |
|---|---|
| codebase_version | v2.1 |
| robot_type | omy_f3m |
| total_episodes | 25 |
| total_frames | 19149 |
| total_tasks | 1 |
| total_videos | 75 |
| total_chunks | 1 |
| chunks_size | 1000 |
| fps | 30 |
| splits | train: 0:25 |
| data_path | data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet |
| video_path | videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4 |
特征 (Features)
| 特征名 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
| timestamp | float32 | [1] | 时间戳 |
| frame_index | int64 | [1] | 帧索引 |
| episode_index | int64 | [1] | 片段索引 |
| index | int64 | [1] | 索引 |
| task_index | int64 | [1] | 任务索引 |
| observation.images.cam_wrist | video | [480, 848, 3] | 腕部摄像头图像 (480x848, 3通道, 编码: libx264, 像素格式: yuv420p) |
| observation.images.cam_top | video | [480, 640, 3] | 顶部摄像头图像 (480x640, 3通道, 编码: libx264, 像素格式: yuv420p) |
| observation.images.cam_side | video | [480, 640, 3] | 侧面摄像头图像 (480x640, 3通道, 编码: libx264, 像素格式: yuv420p) |
| observation.state | float32 | [7] | 机器人状态 (关节1-6, rh_r1_joint) |
| action | float32 | [7] | 动作 (关节1-6, rh_r1_joint) |
关键信息
- 数据集包含 25 个片段 (episodes),共计 19149 帧。
- 所有片段均用于 训练 拆分。
- 机器人类型为 omy_f3m。
- 仅包含 1 个任务。
- 提供 3 个摄像头视角:腕部 (cam_wrist)、顶部 (cam_top)、侧面 (cam_side)。
- 状态和动作均为 7 维,对应 6 个关节加上一个夹爪关节 (rh_r1_joint)。
- 数据以 Parquet 格式存储,视频以 MP4 格式存储。



