Adversarial Robustness via Random Projection Filters

Environment

• torch 1.7.1
• torchvision 0.8.2
• torchattacks 3.2.6

Training of RPF

• To train a ResNet18 with RPF on CIFAR-10:

python train.py --network ResNet18 --dataset cifar10 --attack_iters 10 --lr_schedule multistep --epochs 200 --adv_training --rp --rp_block -1 -1 --rp_out_channel 48 --rp_weight_decay 1e-2 --save_dir resnet18_c10_RPF

• To train a ResNet50 with RPF on ImageNet:

python train_imagenet.py --pretrained --lr 0.02 --lr_schedule cosine --batch_size 1024 --epochs 90 --adv_train --rp --rp_block -1 -1 --rp_out_channel 48 --rp_weight_decay 1e-2 --save_dir resnet50_imagenet_RPF

Evaluation of RPF

• To evaluate the performance of ResNet18 with RPF on CIFAR-10:

python evaluate.py --dataset cifar10 --network ResNet18 --rp --rp_out_channel 48 --rp_block -1 -1 --save_dir eval_r18_c10 --pretrain [path_to_model]

• To evaluate the performance of ResNet50 with RPF on ImageNet:

python train_imagenet.py --evaluate --rp --rp_out_channel 48 --save_dir eval_r50_imagenet --eval_model_path [path_to_model]

Pretrained Models

Pretrained models are provided in google-drive.