rollout_pick_n_place_dagger_r8
收藏资源简介:
该数据集由LeRobot创建,专注于机器人学任务,特别是拾放(pick and place)操作。数据集包含机器人控制所需的多模态数据:动作指令(action)为6维浮点数组,对应肩部平移、肩部抬升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置;观测状态(observation.state)为6维浮点数组,反映机器人关节位置;观测图像包括前端摄像头图像(observation.images.front)和标注图像(observation.images.annotated),均为480x640分辨率、3通道的RGB视频,帧率30 fps,编码为av1。此外,数据集还包含时间戳、帧索引、回合索引、索引和任务索引等元数据。数据规模为1个回合、1870帧、1个任务,数据文件总大小100 MB,视频文件总大小200 MB。数据以Parquet和MP4格式存储,适用于机器人模仿学习、视觉运动控制等任务,支持训练集划分。机器人类型为so_follower。
This dataset was created by LeRobot, focusing on robotics tasks, particularly pick-and-place operations. It contains multimodal data required for robot control: the action command (action) is a 6-dimensional float array corresponding to shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist rotation and gripper position; the observation state (observation.state) is a 6-dimensional float array that reflects the robot's joint positions. The observation images include front-facing camera images (observation.images.front) and annotated images (observation.images.annotated), both are 480x640 resolution, 3-channel RGB videos with a frame rate of 30 fps and encoded in AV1. In addition, the dataset also includes metadata such as timestamps, frame indices, episode indices, indices and task indices. The dataset has a scale of 1 episode, 1870 frames and 1 task, with a total size of 100 MB for data files and 200 MB for video files. The data is stored in Parquet and MP4 formats, suitable for tasks such as robot imitation learning and visual motion control, and supports training set splitting. The robot type is so_follower.
数据集概述
- 数据集名称:rollout_pick_n_place_dagger_r8
- 任务类别:机器人(robotics)
- 许可证:Apache-2.0
- 创建工具:使用 LeRobot 创建
数据集结构
- 数据格式:parquet 文件,路径为
data/*/*.parquet - 帧率:30 FPS
- 机器人类型:so_follower
特征(Features)
| 特征名称 | 数据类型 | 形状 | 描述 |
|---|---|---|---|
action |
float32 | [6] | 机器人动作:shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos |
observation.state |
float32 | [6] | 机器人状态:shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos |
observation.images.front |
video | [480, 640, 3] | 前视摄像头视频(AV1编码,30 FPS,640x480,RGB) |
observation.images.annotated |
video | [480, 640, 3] | 标注版摄像头视频(AV1编码,30 FPS,640x480,RGB) |
timestamp |
float32 | [1] | 时间戳 |
frame_index |
int64 | [1] | 帧索引 |
episode_index |
int64 | [1] | 回合索引 |
index |
int64 | [1] | 全局索引 |
task_index |
int64 | [1] | 任务索引 |
数据集统计
- 总回合数:1
- 总帧数:1870
- 总任务数:1
- 数据分块大小:1000
- 数据文件大小:约 100 MB
- 视频文件大小:约 200 MB
- 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 - 数据划分:训练集(train)包含全部数据(0:1)
可视化
- 可通过 LeRobot 可视化工具 在线浏览数据集。




