oxe-auge/fractal20220817_data_train_67500_70000_augmented

fractal20220817_data_train_67500_70000_augmented是一个机器人学数据集，属于OXE-AugE项目的一部分。该数据集专注于机器人数据的增强，用于跨体现策略学习。它包含多种机器人的增强图像和运动学数据，以及自然语言指令和片段元数据。数据集采用CC-BY-4.0许可，适用于研究目的。数据集包含2500个片段，108,790帧，帧率为3FPS，支持多种机器人型号（如jaco、kinova3、kuka_iiwa等）。数据以parquet文件格式存储，包含丰富的特征如机器人图像、末端执行器姿态、关节位置等。

The fractal20220817_data_train_67500_70000_augmented is a robotics dataset part of the OXE-AugE project, focusing on augmenting robot data for cross-embodiment policy learning. It includes augmented images and kinematic data for various robots, along with natural language instructions and episode metadata. Licensed under CC-BY-4.0, its designed for research purposes. The dataset contains 2,500 episodes, 108,790 frames at 3 FPS, and supports multiple robot models (e.g., jaco, kinova3, kuka_iiwa). Data is stored in parquet files with rich features like robot images, end-effector poses, joint positions, etc.