Silhouettes

该数据集包含了来自101个类别的物体剪影，旨在用于无监督学习和生成建模任务。通过这些数据，研究人员可以探索在不依赖标注信息的情况下，如何让机器学习模型自动学习和生成物体的轮廓和结构。

This dataset comprises object silhouettes across 101 distinct categories, and is intended for unsupervised learning and generative modeling tasks. Using this dataset, researchers can investigate how machine learning models can automatically learn and generate the contours and structures of objects without relying on labeled information.