oxe-aug/language_table_train_305000_310000_augmented

language_table_train_305000_310000_augmented数据集是OXE-AugE项目的一部分，包含从305000到310000的增强语言表格训练数据。这个数据集基于v3.0代码库版本，包含了多种机器人类型，如google_robot、jaco、kinova3等。数据集以10帧每秒的帧率记录，共有5000个剧集，总共有80189帧。每个剧集包含增强的机器人图像、自然语言指令、机器人末端执行器位置、机器人关节角度等信息。数据集支持跨机器人政策学习的扩展。

The language_table_train_305000_310000_augmented dataset is part of the OXE-AugE project, containing augmented language table training data from 305000 to 310000. This dataset is based on the v3.0 codebase version and includes various robot types such as google_robot, jaco, kinova3, etc. The dataset is recorded at a frame rate of 10 frames per second, with a total of 5,000 episodes and 80,189 frames. Each episode includes augmented robot images, natural language commands, robot end-effector positions, robot joint angles, and other information. The dataset supports the scaling of cross-embodiment policy learning.