oxe-auge/fractal20220817_data_train_30000_32500_augmented

fractal20220817_data_train_30000_32500_augmented是一个用于机器人技术研究的大规模数据集，包含多种机器人（jaco, kinova3, kuka_iiwa, panda, sawyer, ur5e, xarm7）的图像和运动数据。数据集包含2500个episodes和110,276帧，以3帧每秒的速率采集。数据以parquet格式存储，路径为data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet。特征包括机器人图像（256×320×3）、关节位置、末端执行器位置、自然语言指令等。该数据集是OXE-AugE项目的一部分，旨在支持跨机器人策略学习的研究。

fractal20220817_data_train_30000_32500_augmented is a large-scale dataset for robotics research, containing images and motion data of multiple robots (jaco, kinova3, kuka_iiwa, panda, sawyer, ur5e, xarm7). The dataset includes 2,500 episodes and 110,276 frames, captured at 3 frames per second. Data is stored in parquet format at data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet. Features include robot images (256×320×3), joint positions, end-effector poses, natural language instructions, etc. This dataset is part of the OXE-AugE project, aiming to support research in cross-embodiment policy learning.