so101_pick_toy_v15_pre_20260703_185621

该数据集由LeRobot创建，专注于机器人技术任务，具体涉及机器人操作。数据集包含一个总时长为524帧的单个训练集，数据以Parquet文件格式存储，并包含视频文件。特征包括机器人动作（如肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部滚动和夹爪位置）、状态观测（与动作相同的关节位置）、来自腕部和顶部摄像头的图像观测（480x640分辨率，RGB通道，30fps），以及时间戳、帧索引、任务索引等元数据。机器人类型为“so_follower”，数据集采用Apache 2.0许可证，适用于机器人控制和视觉任务的研究。

This dataset was created by LeRobot, focusing on robotics tasks with a particular emphasis on robotic manipulation. It comprises a single training set totaling 524 frames, with data stored in Parquet file format and accompanied by video files. The included features are robot actions (such as shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist roll, and gripper position), state observations (consisting of the same joint positions as the actions), image observations captured by the wrist and top-mounted cameras (with a resolution of 480×640, RGB channels, and 30 frames per second), as well as metadata such as timestamps, frame indices, and task indices. The robotic platform type is "so_follower". This dataset is licensed under the Apache 2.0 License and is suitable for research in robotic control and vision tasks.

原始信息汇总

数据集概述

• 名称: kamada-riku/so101_pick_toy_v15_pre_20260703_185621
• 许可证: Apache-2.0
• 任务类别: 机器人学 (Robotics)
• 标签: LeRobot

数据集来源与工具

• 该数据集使用 LeRobot 创建。

数据集结构

• 总集数 (Episodes): 1
• 总帧数 (Frames): 524
• 总任务数 (Tasks): 1
• 数据文件大小: 100 MB
• 视频文件大小: 200 MB
• 帧率 (FPS): 30
• 机器人类型: so_follower

特征 (Features)

特征名称	数据类型	形状	说明
action	float32	[6]	动作：6个机器人关节位置（shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper）
observation.state	float32	[6]	观测状态：6个机器人关节位置（与动作特征同名）
observation.images.wrist	video	[480, 640, 3]	手腕相机图像：高度480，宽度640，3通道彩色，AV1编码，30 FPS
observation.images.top	video	[480, 640, 3]	顶部相机图像：高度480，宽度640，3通道彩色，AV1编码，30 FPS
timestamp	float32	[1]	时间戳
frame_index	int64	[1]	帧索引
episode_index	int64	[1]	集索引
index	int64	[1]	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
• 分块大小 (Chunks size): 1000

数据划分

• 训练集 (train): 0:1（仅一个划分，全部数据用于训练）