Pytorch implementation of "Adversarial Vertex Mixup: Toward Better Adversarially Robust Generalization"
Official implementation (tensorflow): https://github.com/Saehyung-Lee/cifar10_challenge
Paper: https://arxiv.org/abs/2003.02484
- Network: WideResNet32-10
- Batch size: 128
- Num of iters: 80000 iterations(about 200 epochs)
- Dataset: CIFAR-10
- Optimizer (learning rate): Momentum SGD (0.1)
| Clean | FGSM | PGD-10 | PGD-20 | CW-20 | |
|---|---|---|---|---|---|
| Paper results | 93.24 | 78.25 | 62.67 | 58.23 | 53.63 |
| Impremented results | 94.97 | 83.69 | 62.63 | 57.54 | 40.5 |