droid-30-converted

该数据集是一个机器人数据集，使用LeRobot框架创建。它包含30个剧集，总计7726帧，覆盖27个不同的任务。数据集基于熊猫机器人（panda）设计，数据以Parquet文件格式存储，并包含MP4格式的视频文件。特征数据包括：三个图像特征（exterior_image_1_left、exterior_image_2_left、wrist_image_left），每个图像的尺寸为180x320x3，帧率为15fps；关节位置（joint_position），为7维浮点数组；夹爪位置（gripper_position），为1维浮点数组；动作（actions），为8维浮点数组；以及时间戳、帧索引、剧集索引、索引和任务索引等元数据。数据集总大小为100MB，视频文件总大小为200MB，适用于机器人学习和控制任务，如运动规划、强化学习和模仿学习。This dataset is a robotics dataset created using the LeRobot framework. It contains 30 episodes, totaling 7726 frames, covering 27 different tasks. The dataset is designed based on the Panda robot, with data stored in Parquet file format and includes video files in MP4 format. Feature data includes: three image features (exterior_image_1_left, exterior_image_2_left, wrist_image_left), each with dimensions of 180x320x3 and a frame rate of 15fps; joint positions (joint_position), as a 7-dimensional floating-point array; gripper positions (gripper_position), as a 1-dimensional floating-point array; actions (actions), as an 8-dimensional floating-point array; and metadata such as timestamps, frame indices, episode indices, indices, and task indices. The total dataset size is 100MB, with video files totaling 200MB, suitable for robotics learning and control tasks, such as motion planning, reinforcement learning, and imitation learning.