Dataset Condensation with Differentiable Siamese Augmentation
In many machine learning problems, large-scale datasets have become the de-facto standard to train state-of-the-art deep networks at the price of heavy computation load. In this paper, we focus on...
arXiv.org
