robot-learning-team43/so101_filtered_dohyung_HQ_molmoact

该数据集是一个机器人控制数据集，使用LeRobot创建。数据集包含机器人动作和观测状态，具体特征包括：6维浮点型动作向量（对应肩部平移、肩部提升、肘部弯曲、腕部弯曲、腕部旋转和夹爪位置），6维浮点型观测状态向量（与动作相同维度），以及前视摄像头图像（视频帧，分辨率为480x640，3通道RGB，帧率30fps）。此外，还包括时间戳、帧索引、episode索引、索引和任务索引等元数据。数据集总共有143个episodes，152,784帧，数据文件总大小为100MB，视频文件总大小为200MB，主要用于机器人学习和控制任务。This dataset is a robotics control dataset created using LeRobot. It includes robot actions and observation states with features such as: a 6-dimensional float32 action vector (corresponding to shoulder pan, shoulder lift, elbow flex, wrist flex, wrist roll, and gripper positions), a 6-dimensional float32 observation state vector (same dimensions as actions), and front-view camera images (video frames with 480x640 resolution, 3 RGB channels, 30fps frame rate). Additionally, it contains metadata like timestamp, frame index, episode index, index, and task index. The dataset consists of 143 episodes, 152,784 frames, with a total data file size of 100MB and video file size of 200MB, primarily used for robot learning and control tasks.