oxe-aug/fractal20220817_data_train_20000_30000_augmented

fractal20220817_data_train_20000_30000_augmented数据集是基于原始数据集进行增强的大型机器人数据集，包含kinova3、kuka_iiwa、panda、sawyer、ur5e、xarm7等多种机器人类型，总共10,000个剧集，431,028个帧，70,000个视频。数据集以3帧每秒的帧率记录，并提供了包括机器人末端执行器位置、关节角度、图像等多种特征。该数据集适用于机器人学的增强学习和交叉身体策略学习。

The fractal20220817_data_train_20000_30000_augmented dataset is a large-scale robot augmentation dataset based on the original dataset, containing various robot types such as kinova3, kuka_iiwa, panda, sawyer, ur5e, xarm7, with a total of 10,000 episodes, 431,028 frames, and 70,000 videos. The dataset is recorded at a frame rate of 3 frames per second and provides various features including robot end-effector positions, joint angles, images, etc. This dataset is suitable for reinforcement learning and cross-embodiment policy learning in robotics.