Code for our paper "Regularizing Neural Networks via Adversarial Model Perturbation", CVPR2021 - View it on GitHub