danielsanjosepro/lego_stack_f10_v1

这是一个关于机器人学的数据集，包含30个剧集，共5246帧，专注于1个任务。数据集由60个视频组成，每个视频被分割成1个片段，每个片段包含1000帧。数据集的特征包括主要观测图像、手腕观测图像、笛卡尔状态、抓手状态、关节状态、外部力传感器状态、目标状态、动作、时间戳、帧索引、剧集索引、索引和任务索引。所有视频的帧率为10fps，且视频格式为av1编码的yuv420p，没有音频。

This is a robotics dataset containing 30 episodes, totaling 5246 frames, focused on 1 task. The dataset consists of 60 videos, each divided into 1 chunk with 1000 frames. The dataset features include primary observation images, wrist observation images, Cartesian state, gripper state, joint state, external wrench sensor state, target state, action, timestamp, frame index, episode index, index, and task index. All videos are at 10fps and formatted in av1 encoded yuv420p without audio.