oxe-auge/fractal20220817_data_train_57500_60000_augmented

fractal20220817_data_train_57500_60000_augmented是一个用于机器人学习和增强学习的数据集。数据集包含多个机器人（如jaco、kinova3、kuka_iiwa、panda、sawyer、ur5e、xarm7）的图像、关节位置、末端执行器位置等信息，并且提供了自然语言指令。数据集的结构包括训练集，总共有2,500个episodes和108,036帧，帧率为3fps。数据以parquet文件格式存储，每个文件包含多个特征，如机器人图像、关节角度、末端执行器位置等。数据集还提供了详细的特征描述和引用政策。

fractal20220817_data_train_57500_60000_augmented is a dataset for robotics and reinforcement learning. It includes images, joint positions, end-effector positions, and natural language instructions for multiple robots (e.g., jaco, kinova3, kuka_iiwa, panda, sawyer, ur5e, xarm7). The dataset is structured with a training split, containing 2,500 episodes and 108,036 frames at 3fps. Data is stored in parquet files, with each file containing various features such as robot images, joint angles, and end-effector positions. The README provides detailed feature descriptions and citation policies.