FunnyBirds

该数据集名为FunnyBirds基准，它涵盖了用于评估原型部分网络中解释各方面的一系列指标。此外，该框架采用完整性、正确性和对比性等指标来评估解释的效能。该数据集的任务是对不完整的鸟类图像进行多标签分类。

This dataset, named the FunnyBirds benchmark, encompasses a suite of metrics designed to evaluate diverse aspects of explanations in prototypical part networks. Furthermore, this framework employs metrics such as completeness, correctness, and contrastiveness to assess the efficacy of explanations. The core task of this dataset is multi-label classification for incomplete bird images.