aboardman/record-test_20260529_122057
收藏资源简介:
该数据集是一个用于机器人任务的数据集,使用LeRobot创建。数据集包含一个episode,总共有1728帧,帧率为30fps。特征包括动作(action)和观测(observation),其中动作由6个浮点数值组成,代表机器人关节位置(如肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置)。观测包括状态(state)和图像(images),状态同样由6个浮点数值表示关节位置;图像分为前视(front)和俯视(overhead)两种视角,均为视频格式,分辨率480x640,3通道(RGB),使用AV1编解码器。此外,数据集还包含时间戳、帧索引、episode索引、索引和任务索引等元数据。机器人类型为so_follower,数据以parquet文件格式存储,总数据文件大小为100MB,视频文件大小为200MB。数据集仅包含训练分割(train split)。
This is a robotics task dataset developed using LeRobot. It consists of one episode with a total of 1728 frames at a frame rate of 30 fps. The dataset includes two core feature categories: action and observation. The action is composed of 6 floating-point values, representing the robot's joint positions including shoulder translation, shoulder elevation, elbow flexion, wrist flexion, wrist rotation, and gripper position. The observation comprises two sub-components: state and images. The state is also represented by 6 floating-point values corresponding to the robot's joint positions. The images are available in two perspectives: front view and overhead view, both in video format with a resolution of 480 × 640, 3-channel RGB, and encoded using the AV1 codec. Additionally, the dataset contains various metadata fields such as timestamps, frame indices, episode indices, sample indices, and task indices. The robot platform for this dataset is so_follower. The data is stored in Parquet file format, with a total data file size of 100 MB and a video file size of 200 MB. The dataset only includes the training split.




