record-test_20260603_172830

该数据集是使用LeRobot框架创建的机器人学习数据集，包含10个训练集片段，总计17678帧数据。数据集结构包括：6维浮点型动作向量（控制肩部平移、肩部抬升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置）；观测状态包含相同的6维关节位置；视觉观测包含顶部摄像头和腕部摄像头的视频数据（分辨率480x640，30fps，AV1编码）；以及时间戳、帧索引、任务索引等元数据。数据以Parquet格式存储，视频单独存储为MP4格式，总数据量约300MB。数据集针对follow任务设计，使用Apache 2.0许可证。This is a robotic learning dataset created using the LeRobot framework, consisting of 10 training episodes with a total of 17678 frames. The dataset structure includes: a 6-dimensional floating-point action vector that controls shoulder translation, shoulder elevation, elbow flexion, wrist flexion, wrist rotation, and gripper position; the observation state contains the same 6-dimensional joint positions; visual observations comprise video data from the overhead camera and wrist-mounted camera, with specifications of 480x640 resolution, 30 frames per second (fps), and AV1 encoding; as well as metadata such as timestamps, frame indices, and task indices. The dataset is stored in Parquet format, with videos saved separately in MP4 format, and the total data volume is approximately 300 MB. This dataset is designed for the "follow" task and is licensed under the Apache 2.0 license.

原始信息汇总

数据集概述

• 数据集名称：record-test_20260603_172830
• 许可证：Apache-2.0
• 任务类别：机器人学（Robotics）
• 标签：LeRobot

数据集结构

• 数据格式：Parquet 文件，路径为 data/*/*.parquet
• 帧率：30 FPS
• 总片段数：10
• 总帧数：17678
• 总任务数：1
• 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]	6个关节动作：shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos
observation.state	float32	[6]	6个关节状态，与动作特征名称相同
observation.images.top	video	[480, 640, 3]	顶部摄像头视频，分辨率480x640，AV1编码，30 FPS
observation.images.wrist	video	[480, 640, 3]	腕部摄像头视频，分辨率480x640，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]	索引
task_index	int64	[1]	任务索引

数据划分

• 训练集：包含全部10个片段（索引 0:10）

文件大小

• 数据文件：约 100 MB
• 视频文件：约 200 MB

创建工具

该数据集使用 LeRobot 创建。