Alkatt/eval_LAVLA_Async_VI_100M_40k_b20_whiteblackcube_v2_2026_05_27_white_latency_bench_v1
收藏资源简介:
该数据集是使用LeRobot创建的,专注于机器人技术领域。数据集包含机器人动作、状态观测和图像数据,具体特征包括:动作数据(如肩部平移、肩部升降、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置)、观测状态(与动作数据类似)、以及来自三个摄像头的图像观测(每个摄像头分辨率为480x640,3通道,视频编码为av1,帧率为30fps)。数据集结构包括总episodes为2,总帧数为3394,总任务数为1,数据文件大小为100MB,视频文件大小为200MB,帧率为30fps,分割为训练集(索引0到2)。机器人类型为so_follower,适用于机器人跟随任务。
This dataset was created using LeRobot. It is focused on robotics and includes data on robot actions, state observations, and image observations. Key features encompass action data (e.g., shoulder pan position, shoulder lift position, elbow flex position, wrist flex position, wrist roll position, gripper position), observation state (similar to action data), and image observations from three cameras (each with a resolution of 480x640, 3 channels, video codec av1, and frame rate of 30fps). The dataset structure comprises total episodes of 2, total frames of 3394, total tasks of 1, data file size of 100MB, video file size of 200MB, frame rate of 30fps, and a split for training (indices 0 to 2). The robot type is so_follower, suitable for robot following tasks.




