Mechanichris/Dagger_plus_chris_reps_05192026

该数据集是一个机器人学习数据集，使用LeRobot工具创建，包含双臂机器人的动作和状态观测数据。数据集中有745个总剧集和1,916,069个总帧数，帧率为30 FPS。特征包括16维的动作向量（如左右肩、肘、腕和夹爪的位置和速度）和状态观测向量（相同维度）。此外，数据集提供四个视角的图像数据：前左、前右、腕左和腕右，每个图像分辨率为240x320，3通道彩色视频，使用AV1编解码器。元数据包括时间戳、帧索引、剧集索引、任务索引等。数据集格式为Parquet和MP4视频文件，总数据大小约为100 MB，视频文件大小约为200 MB。机器人类型为sourccey_client，仅包含训练分割。该数据集适用于机器人控制、感知和强化学习任务。This is a robotic learning dataset created using the LeRobot toolkit, containing action and state observation data for dual-arm robots. It includes 745 total episodes and 1,916,069 total frames, with a frame rate of 30 FPS. The features consist of a 16-dimensional action vector (covering positions and velocities of the left and right shoulders, elbows, wrists, and grippers, for example) and a state observation vector of the same dimension. Additionally, the dataset provides image data from four viewpoints: front-left, front-right, wrist-left, and wrist-right. Each is a 3-channel color video with a resolution of 240×320, encoded using the AV1 codec. Metadata includes timestamps, frame indices, episode indices, task indices, and other relevant information. The dataset is provided in Parquet and MP4 video file formats, with an overall dataset size of approximately 100 MB and a total video file size of around 200 MB. The target robot type is sourcery_client, and only the training split is included. This dataset is suitable for robotic control, perception, and reinforcement learning tasks.