yukamatsumae/ToyCarDemoSlow_may_20260527_010435
收藏资源简介:
该数据集由LeRobot创建,专注于机器人技术任务,特别是针对Otter Follower机器人类型。数据集包含5个总集,1500个总帧,以及1个总任务。数据以20帧每秒(fps)的速率采集,并存储在Parquet文件中,总数据文件大小为100MB,视频文件大小为200MB。数据集的特征包括动作(action)和观察状态(observation.state),两者均为6维浮点数组,分别代表机器人关节位置(如肩部平移、肩部提升、肘部弯曲、手腕弯曲、手腕滚动和夹爪位置)。此外,观察图像(observation.images.right)为视频格式,分辨率为480x640,3通道,使用AV1编解码器,无音频。其他特征包括时间戳、帧索引、集索引、索引和任务索引。数据集结构支持分块存储,数据路径和视频路径分别指定。数据集已划分为训练集(集0到4),适用于机器人控制、视觉导航和机器学习任务。
This dataset was developed by LeRobot, focusing on robotics tasks, specifically targeting the Otter Follower robot platform. It contains 5 total splits, 1500 total frames, and 1 single task. The data is collected at a rate of 20 frames per second (fps), stored in Parquet files, with a total data file size of 100MB and video file size of 200MB. The dataset features two core attributes: action and observation.state, both 6-dimensional floating-point arrays that respectively represent robot joint positions including shoulder translation, shoulder lift, elbow flexion, wrist flexion, wrist roll, and gripper position. Additionally, observation.images.right is a video format with a resolution of 480x640, 3 channels, using the AV1 codec and no audio. Other attributes include timestamp, frame index, split index, sample index, and task index. The dataset structure supports chunked storage, with separate specifications for data paths and video paths. The dataset has been partitioned into training splits (split 0 to 4), which is applicable to robotics control, visual navigation, and machine learning tasks.




