AdvDetect/generate_adv.py at main · xinwong/AdvDetect · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
import argparse
import numpy as np
from common.util import *
from setup_paths import *
from art.attacks.evasion import FastGradientMethod, BasicIterativeMethod, CarliniL2Method, CarliniLInfMethod, ProjectedGradientDescent, DeepFool, SpatialTransformation, SquareAttack, ZooAttack, AdversarialPatchPyTorch

def main(args):
    set_seed(args)

    assert args.dataset in ['mnist', 'cifar', 'svhn', 'imagenet'], \
        "dataset parameter must be either 'mnist', 'cifar', or 'imagenet'"
    print('Dataset: %s' % args.dataset)

    if args.dataset == 'mnist':
        from baseline.cnn.cnn_mnist import MNISTCNN as model
        model_mnist = model(mode='load', filename='cnn_{}.pt'.format(args.dataset))
        classifier = model_mnist.classifier
        epsilons=[8/256, 16/256, 32/256, 64/256, 80/256, 128/256]
        epsilons1=[5, 10, 15, 20, 25, 30, 40]
        epsilons2=[0.125, 0.25, 0.3125, 0.5, 1, 1.5, 2]
        eps_sa=0.3
        x_test = model_mnist.x_test
        y_test = model_mnist.y_test
        translation = 10
        rotation = 60

    elif args.dataset == 'cifar':
        from baseline.cnn.cnn_cifar10 import CIFAR10CNN as model
        model_cifar = model(mode='load', filename='cnn_{}.pt'.format(args.dataset))
        classifier = model_cifar.classifier
        epsilons=[8/256, 16/256, 32/256, 64/256, 80/256, 128/256]
        epsilons1=[5, 10, 15, 20, 25, 30, 40]
        epsilons2=[0.125, 0.25, 0.3125, 0.5, 1, 1.5, 2]
        eps_sa=0.125
        x_test = model_cifar.x_test
        y_test = model_cifar.y_test
        translation = 8
        rotation = 30

    elif args.dataset == 'svhn':
        from baseline.cnn.cnn_svhn import SVHNCNN as model
        model_svhn = model(mode='load', filename='cnn_{}.pt'.format(args.dataset))
        classifier = model_svhn.classifier
        epsilons=[8/256, 16/256, 32/256, 64/256, 80/256, 128/256]
        epsilons1=[5, 10, 15, 20, 25, 30, 40]
        epsilons2=[0.125, 0.25, 0.3125, 0.5, 1, 1.5, 2]
        eps_sa=0.125
        x_test = model_svhn.x_test
        y_test = model_svhn.y_test
        translation = 10
        rotation = 60

    elif args.dataset == 'imagenet':
        from baseline.cnn.cnn_imagenet import ImageNetCNN as model
        model_imagenet = model(filename='cnn_{}.pt'.format(args.dataset))
        classifier = model_imagenet.classifier
        epsilons=[8/256, 16/256, 32/256, 64/256, 80/256, 128/256]
        epsilons1=[5, 10, 15, 20, 25, 30, 40]
        epsilons2=[0.125, 0.25, 0.3125, 0.5, 1, 1.5, 2]
        eps_sa=0.125
        x_test = model_imagenet.x_val
        y_test = model_imagenet.y_val
        translation = 8
        rotation = 30

    # #FGSM
    # for e in epsilons:
    #     attack = FastGradientMethod(estimator=classifier, eps=e, eps_step=0.01, batch_size=256)
    #     adv_data = attack.generate(x=x_test)
    #     adv_file_path = adv_data_dir + args.dataset + '_fgsm_' + str(e) + '.npy'
    #     np.save(adv_file_path, adv_data)
    #     print('Done - {}'.format(adv_file_path))

    # #BIM
    # for e in epsilons:
    #     attack = BasicIterativeMethod(estimator=classifier, eps=e, eps_step=0.01, batch_size=32, max_iter=int(e*256*1.25))
    #     adv_data = attack.generate(x=x_test)
    #     adv_file_path = adv_data_dir + args.dataset + '_bim_' + str(e) + '.npy'
    #     np.save(adv_file_path, adv_data)
    #     print('Done - {}'.format(adv_file_path))

    # #PGD1
    # for e in epsilons1:
    #     attack = ProjectedGradientDescent(estimator=classifier, norm=1, eps=e, eps_step=4, batch_size=32)
    #     adv_data = attack.generate(x=x_test)
    #     adv_file_path = adv_data_dir + args.dataset + '_pgd1_' + str(e) + '.npy'
    #     np.save(adv_file_path, adv_data)
    #     print('Done - {}'.format(adv_file_path))

    # #PGD2
    # for e in epsilons2:
    #     attack = ProjectedGradientDescent(estimator=classifier, norm=2, eps=e, eps_step=0.1, batch_size=32)
    #     adv_data = attack.generate(x=x_test)
    #     adv_file_path = adv_data_dir + args.dataset + '_pgd2_' + str(e) + '.npy'
    #     np.save(adv_file_path, adv_data)
    #     print('Done - {}'.format(adv_file_path))

    # #PGDInf
    # for e in epsilons:
    #     attack = ProjectedGradientDescent(estimator=classifier, norm=np.inf, eps=e, eps_step=0.01, batch_size=32)
    #     adv_data = attack.generate(x=x_test)
    #     adv_file_path = adv_data_dir + args.dataset + '_pgdi_' + str(e) + '.npy'
    #     np.save(adv_file_path, adv_data)
    #     print('Done - {}'.format(adv_file_path))

    # #CWi
    # attack = CarliniLInfMethod(classifier=classifier, max_iter=20)
    # adv_data = attack.generate(x=x_test)
    # adv_file_path = adv_data_dir + args.dataset + '_cwi.npy'
    # np.save(adv_file_path, adv_data)
    # print('Done - {}'.format(adv_file_path))

    # #CW2 - SLOW
    # attack = CarliniL2Method(classifier=classifier, max_iter=10, confidence=10)
    # adv_data = attack.generate(x=x_test)
    # adv_file_path = adv_data_dir + args.dataset + '_cw2.npy'
    # np.save(adv_file_path, adv_data)
    # print('Done - {}'.format(adv_file_path))

    #DF
    attack = DeepFool(classifier=classifier)
    adv_data = attack.generate(x=x_test)
    adv_file_path = adv_data_dir + args.dataset + '_df.npy'
    np.save(adv_file_path, adv_data)
    print('Done - {}'.format(adv_file_path))

    #Spatial transofrmation attack
    attack = SpatialTransformation(classifier=classifier, max_translation=translation, max_rotation=rotation)
    adv_data = attack.generate(x=x_test)
    adv_file_path = adv_data_dir + args.dataset + '_sta.npy'
    np.save(adv_file_path, adv_data)
    print('Done - {}'.format(adv_file_path))

    #Square Attack
    attack = SquareAttack(estimator=classifier, max_iter=200, eps=eps_sa)
    adv_data = attack.generate(x=x_test, y=y_test)
    adv_file_path = adv_data_dir + args.dataset + '_sa.npy'
    np.save(adv_file_path, adv_data)
    print('Done - {}'.format(adv_file_path))

    #ZOO attack
    # attack = ZooAttack(classifier=classifier, batch_size=32)
    # adv_data = attack.generate(x=x_test, y=y_test)
    # adv_file_path = adv_data_dir + args.dataset + '_zoo.npy'
    # np.save(adv_file_path, adv_data)
    # print('Done - {}'.format(adv_file_path))

    #Adversarial Patch attack
    if args.dataset != 'mnist':
        attack = AdversarialPatchPyTorch(
                    estimator=classifier,
                    rotation_max=22.5,
                    scale_min=0.4,
                    scale_max=1.0,
                    learning_rate=1,
                    batch_size=32,
                    max_iter=1000,
                    patch_shape=(3, 100, 100),
                    verbose=True,
                    optimizer='pgd'
                )
        patch, patch_mask= attack.generate(x=x_test, y=y_test)
        save_image(torch.from_numpy(patch * patch_mask), 'ap.jpg')
        adv_data = attack.apply_patch(x=x_test, scale=0.2)
        adv_file_path = adv_data_dir + args.dataset + '_ap.npy'
        np.save(adv_file_path, adv_data)
        print('Done - {}'.format(adv_file_path))

if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument('-d', '--dataset', required=True, type=str, help="Dataset to use; either 'mnist', 'cifar', or 'imagenet'")
    parser.add_argument('-s', '--seed', help='set seed for model', default=123, type=int)
    args = parser.parse_args()
    main(args)