pick_place_v4
收藏资源简介:
该数据集使用LeRobot框架创建,是一个机器人操作数据集,专门设计用于单一任务场景。它包含102个训练片段,总计37,005帧数据。数据采用结构化存储,包括动作指令、机器人状态观测和多视角视觉观测。动作和状态观测均包含6个浮点数维度,对应机械臂的肩部平移、肩部抬升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置。视觉观测提供两个视角:腕部摄像头(分辨率480×640)和全局概览摄像头(分辨率720×1280),均以30fps的视频格式存储。此外,数据集还包含时间戳、帧索引、片段索引、数据索引和任务索引等元数据字段。该数据集适用于机器人模仿学习、行为克隆、离线强化学习等任务,为机器人操作策略的训练与评估提供了多模态交互数据。
This dataset is created using the LeRobot framework and is a robotic manipulation dataset designed for single-task scenarios. It contains 102 training episodes, totaling 37,005 frames of data. The data is stored in a structured format, including action commands, robot state observations, and multi-view visual observations. Specifically, both actions and state observations consist of 6 floating-point dimensions corresponding to the shoulder translation, shoulder lift, elbow bend, wrist bend, wrist rotation, and gripper position of the robotic arm. Visual observations provide two perspectives: a wrist camera (resolution 480×640) and an overview camera (resolution 720×1280), both stored as 30fps video formats. Additionally, the dataset includes metadata fields such as timestamps, frame indices, episode indices, data indices, and task indices. This dataset is suitable for tasks like robotic imitation learning, behavior cloning, and offline reinforcement learning, providing multimodal interaction data for training and evaluating robotic manipulation policies.
数据集概述
- 数据集名称:
pick_place_v4 - 许可证: Apache-2.0
- 任务类别: 机器人学 (robotics)
- 标签: LeRobot
数据集结构
- 机器人类型:
so_follower - 总片段数 (episodes): 102
- 总帧数 (frames): 37,005
- 任务数量: 1
- 数据分块大小: 1000
- 数据文件总大小: 100 MB
- 视频文件总大小: 200 MB
- 帧率 (FPS): 30
- 数据划分: 训练集 (train) 包含全部 102 个片段
数据特征
| 特征名称 | 数据类型 | 形状 | 描述 |
|---|---|---|---|
action |
float32 | [6] | 动作指令,包含6个关节位置: shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos |
observation.state |
float32 | [6] | 观察到的机器人状态,与动作特征相同的6个关节位置 |
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.overview |
video | [720, 1280, 3] | 全局视角摄像头视频,分辨率 720x1280,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
- 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
引用信息
- 主页: [More Information Needed]
- 论文: [More Information Needed]
- BibTeX: [More Information Needed]




