eval_xvla-so101-4tasks-aug_pick-place_25

该数据集是一个使用LeRobot创建的机器人控制数据集，专门针对so_follower类型的机器人。数据集包含16个训练片段，共计31234帧数据，以parquet格式存储。数据特征包括：6自由度的机器人动作（关节位置）和状态观测（同样为6个关节位置），以及两个来自不同视角的480x640分辨率、30fps的RGB视频观测。此外，还包含时间戳、帧索引、片段索引、任务索引等元数据。数据集主要用于机器人控制相关的训练和评估任务。

This is a robot control dataset created using LeRobot, specifically tailored for so_follower-type robots. The dataset contains 16 training episodes, with a total of 31234 frames of data, stored in Parquet format. The data features include: 6-degree-of-freedom robot actions (joint positions) and state observations (also comprising 6 joint positions), as well as two RGB video observations from different perspectives with a resolution of 480x640 and a frame rate of 30 fps. Additionally, it includes metadata such as timestamps, frame indices, episode indices, and task indices. This dataset is primarily intended for training and evaluation tasks related to robot control.

原始信息汇总

数据集概览

• 数据集名称: eval_xvla-so101-4tasks-aug_pick-place_25
• 许可证: Apache-2.0
• 任务类别: 机器人技术 (Robotics)
• 标签: LeRobot
• 创建工具: 使用 LeRobot 创建

数据集规模与结构

• 机器人类型: so_follower
• 总片段数 (Episodes): 16
• 总帧数 (Frames): 31234
• 总任务数 (Tasks): 1
• 块大小 (Chunks Size): 1000
• 数据文件大小: 100 MB
• 视频文件大小: 200 MB
• 帧率 (FPS): 30
• 数据划分: 全部16个片段用于训练 (train: "0:16")

数据存储路径

• 数据文件: data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet
• 视频文件: videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4

特征 (Features)

特征名称	数据类型	形状	描述
action	float32	[6]	6维动作：shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos
observation.state	float32	[6]	6维机器人状态：shoulder_pan.pos, shoulder_lift.pos, elbow_flex.pos, wrist_flex.pos, wrist_roll.pos, gripper.pos
observation.images.image	video	[480, 640, 3]	第一视角视频，分辨率480x640，3通道，帧率30 FPS，编解码器AV1，像素格式yuv420p，非深度图，无音频
observation.images.image2	video	[480, 640, 3]	第二视角视频，与image规格一致
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

可视化

• 可通过以下空间可视化数据集：Visualize this dataset