so101_pickplace_driverbit_in_cup_20260608_195725
收藏资源简介：
该数据集是一个机器人学数据集，专门用于机器人操作任务，特别是拾取放置（pick-and-place）场景。它由LeRobot平台创建，包含从机器人收集的交互数据，旨在支持机器人学习和控制研究。数据集结构包括动作数据（如关节位置）、观察数据（如机器人状态和图像）以及元数据（如时间戳、帧索引等）。具体特征包括：action（6个浮点数，表示肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部滚动和夹爪位置）、observation.state（与action相同的6个浮点数）、observation.images.wrist和observation.images.top（均为视频数据，分辨率480x640，3通道RGB，帧率30fps，使用AV1编解码器）。数据集总共有2个episodes、1351帧和1个任务，数据以parquet文件格式存储，视频文件以mp4格式存储。机器人类型为so_follower，适用于训练和评估机器人控制算法。
This is a robotics dataset dedicated to robotic manipulation tasks, especially pick-and-place scenarios. It was created by the LeRobot platform and contains interaction data collected from robots, aiming to support research on robot learning and control. The dataset structure includes action data (e.g., joint positions), observation data (e.g., robot states and images), and metadata (e.g., timestamps, frame indices, etc.). Specific features are as follows: action consists of 6 floating-point numbers representing shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist roll and gripper position; observation.state contains the same 6 floating-point numbers as action; observation.images.wrist and observation.images.top are both video data with a resolution of 480x640, 3-channel RGB, 30fps frame rate and encoded using the AV1 codec. In total, the dataset has 2 episodes, 1351 frames and 1 task. The data is stored in parquet file format, while the video files are stored in mp4 format. The robot type is so_follower, which is suitable for training and evaluating robotic control algorithms.
数据集概述
- 数据集名称:
chanwook/so101_pickplace_driverbit_in_cup_20260608_195725 - 任务类别: 机器人学 (robotics)
- 标签: LeRobot
- 许可证: Apache-2.0
数据集来源与创建
- 该数据集使用 LeRobot 创建。
数据集结构
- 数据格式: Parquet 文件，位于
data/*/*.parquet。 - 帧率 (fps): 30
- 总片段数 (total_episodes): 2
- 总帧数 (total_frames): 1351
- 总任务数 (total_tasks): 1
- 机器人类型:
so_follower - 数据划分: 仅包含训练集 (train)，划分范围为
0:2（即所有片段用于训练）。 - 数据大小:
- 数据文件大小: 100 MB
- 视频文件大小: 200 MB
特征 (Features)
数据集中包含以下特征：
| 特征名 | 数据类型 | 形状 | 说明 |
|---|---|---|---|
action |
float32 | [6] | 动作指令，包含6个关节位置：shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper |
observation.state |
float32 | [6] | 观测状态，与动作相同的6个关节位置 |
observation.images.wrist |
video | [480, 640, 3] | 腕部摄像头图像，分辨率480x640，3通道，编码为AV1，帧率30fps |
observation.images.top |
video | [480, 640, 3] | 顶部摄像头图像，分辨率480x640，3通道，编码为AV1，帧率30fps |
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 |
数据存储路径
- 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
引用
目前没有提供 BibTeX 引用信息。
