Towards Mitigating Architecture Overfitting on Distilled Datasets

Xuyang Zhong, Chen Liu

Dataset distillation methods have demonstrated remarkable performance for neural networks trained with very limited training data. However, a significant challenge arises in the fo…