ArtMancer/ORiDa-train-only

该数据集包含图像和相关元数据，划分为事实性(factual_only)和事实对比性(factual_counterfactual)两种类型。提供字段包括组ID、样本ID、图像、真实值图像、边界框、点集、掩膜和结果可视化图像。训练集包含143740个示例，数据大小为约135GB。

The dataset includes images and associated metadata, categorized into factual_only and factual_counterfactual types. It provides fields such as group ID, sample ID, image, ground truth image, bounding box, point set, mask, and result visualization image. The training set contains 143740 examples, with a total size of approximately 135GB.