iantc104/leap_hand_grasp_dataset_rrt_knife_v3

该数据集是一个用于机器人操作或物理交互任务的数据集，包含200,000个训练示例。数据特征包括关节位置（q）、关节掩码（q_mask）、物体姿态（object_pose）、目标位置（target_pos）、目标法向量（target_normal）、接触位置（contact_pos）、接触法向量（contact_normal）、接触链接ID（contact_link_id）和批次索引（batch_index）。这些特征可能用于模拟或实际机器人环境中的抓取、操作或运动规划任务，例如控制机器人手臂与物体交互。数据集总大小为约79.6 MB，仅包含训练分割，适用于机器学习模型训练。

This dataset is intended for robotic manipulation or physical interaction tasks, containing 200,000 training examples. Its data features include joint positions (q), joint masks (q_mask), object poses (object_pose), target positions (target_pos), target normals (target_normal), contact positions (contact_pos), contact normals (contact_normal), contact link IDs (contact_link_id), and batch indices (batch_index). These features can be applied to grasping, manipulation or motion planning tasks in simulated or real robotic environments, such as controlling robotic arms to interact with physical objects. The total size of the dataset is around 79.6 MB, and it only contains the training split, which is applicable for machine learning model training.