Test Object Collections

该数据集包含了用于在各种真实世界场景中测试系统的瓶子、碗和马克杯等物品。值得注意的是，系统中用于训练模型的对象并未包括在这些测试物品中。该数据集的规模涉及41个场景，每个场景中有5个来自对象集合的实例。所涉及的任务主要是人机交互以及指令引导下的抓取任务。

This dataset contains items such as bottles, bowls, and mugs for testing systems across various real-world scenarios. Notably, the objects used for model training in the system are not included among these test items. The dataset spans 41 scenarios, with 5 instances from the object set in each scenario. The tasks involved primarily include human-computer interaction and instruction-guided grasping tasks.