gsethia08/part4-point_and_click_remote

这是一个使用LeRobot创建的机器人学习数据集，专注于手部机器人操作任务。数据集包含2144个训练集片段，总计633,314帧数据，涵盖17个不同任务。数据特征包括：1080x1920分辨率的RGB相机观察、48维手部状态（包含左右手腕和手指的坐标与旋转信息）、相机内外参数、48维动作向量、语言指令、时间戳、帧索引、片段索引和任务索引。数据以30fps采集，存储为parquet格式，适用于机器人模仿学习和强化学习研究。This dataset was created using LeRobot and focuses on robotic hand manipulation tasks for machine learning. It contains 2,144 training episodes with a total of 633,314 frames across 17 different tasks. Features include: 1080x1920 RGB camera observations, 48-dimensional hand state (including coordinates and rotations for both left and right wrists and fingers), camera extrinsic and intrinsic parameters, 48-dimensional action vectors, language instructions, timestamps, frame indices, episode indices, and task indices. Data was collected at 30fps and is stored in parquet format, suitable for robotic imitation learning and reinforcement learning research.