train_triple_fps15_3
收藏资源简介:
这是一个使用LeRobot创建的机器人数据集,专注于UR(Universal Robots)机器人。数据集包含机器人的动作(如关节控制)、观测状态(如关节位置)、多视角图像(包括侧视、腕视和顶视摄像头,分辨率为480x640,3通道)以及时间戳、帧索引、回合索引等元数据。数据以Parquet格式存储,视频以MP4格式存储,帧率为15 fps。该数据集适用于机器人学习、控制策略和视觉感知等研究任务。
This is a robot dataset created using LeRobot, focusing on UR (Universal Robots) robots. The dataset includes robot actions (such as joint control), observation states (e.g., joint positions), multi-view images (from side, wrist, and top cameras, with a resolution of 480x640, 3 channels), as well as metadata such as timestamps, frame indices, and episode indices. Data is stored in Parquet format, and videos are stored in MP4 format with a frame rate of 15 fps. It is suitable for research tasks in robot learning, control strategies, and visual perception.
数据集概述
- 名称:
train_triple_fps15_3 - 来源: Hugging Face Datasets
- 许可证: Apache-2.0
- 任务类型: 机器人学(Robotics)
- 标签: LeRobot
数据集创建
该数据集使用 LeRobot 创建,一个用于机器人学习的开源框架。
数据集结构
数据集基于 meta/info.json 文件定义,关键参数如下:
- codebase_version: v3.0
- robot_type: ur(优傲机器人)
- total_episodes: 0
- total_frames: 0
- total_tasks: 0
- chunks_size: 1000
- data_files_size_in_mb: 100 MB
- video_files_size_in_mb: 200 MB
- fps: 15(帧/秒)
- splits: 未定义(空对象)
- data_path:
data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet - video_path:
videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4
特征(Features)
| 特征名称 | 数据类型 | 形状 | 描述 |
|---|---|---|---|
action |
float32 | (7,) | 机器人动作,包含7个关节:shoulder_pan_joint, shoulder_lift_joint, elbow_joint, wrist_1_joint, wrist_2_joint, wrist_3_joint, gripper_joint |
observation.state |
float32 | (7,) | 机器人状态,与动作相同的7个关节变量 |
observation.images.side |
video | (480, 640, 3) | 侧面视角图像,分辨率480x640,RGB三通道 |
observation.images.wrist |
video | (480, 640, 3) | 手腕视角图像,分辨率480x640,RGB三通道 |
observation.images.top |
video | (480, 640, 3) | 顶部视角图像,分辨率480x640,RGB三通道 |
timestamp |
float32 | (1,) | 时间戳 |
frame_index |
int64 | (1,) | 帧索引 |
episode_index |
int64 | (1,) | 回合索引 |
index |
int64 | (1,) | 索引 |
task_index |
int64 | (1,) | 任务索引 |
配置
- 配置名称:
default - 数据文件:
data/*/*.parquet



