-
Notifications
You must be signed in to change notification settings - Fork 3
/
attack_utils.py
124 lines (104 loc) · 4.27 KB
/
attack_utils.py
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
import torch
import random
import numpy as np
from PIL import Image
import json
import os
import pandas as pd
from foolbox import PyTorchModel
import torchvision.models as models
from datetime import datetime
import pandas as pd
def get_model(args,device):
model_name = args.model_name
if model_name == 'resnet-18':
model = models.resnet18(pretrained=True).eval().to(device)
mean = torch.Tensor([0.485, 0.456, 0.406])
std = torch.Tensor([0.229, 0.224, 0.225])
if torch.cuda.is_available():
mean = mean.to(device)
std = std.to(device)
preprocessing = dict(mean=mean, std=std, axis=-3)
fmodel = PyTorchModel(model, bounds=(0, 1), preprocessing=preprocessing)
return fmodel
elif model_name == 'inception-v3':
model = models.inception_v3(pretrained=True).eval().to(device)
mean = torch.Tensor([0.485, 0.456, 0.406])
std = torch.Tensor([0.229, 0.224, 0.225])
if torch.cuda.is_available():
mean = mean.cuda(0)
std = std.cuda(0)
preprocessing = dict(mean=mean, std=std, axis=-3)
fmodel = PyTorchModel(model, bounds=(0, 1), preprocessing=preprocessing)
return fmodel
elif model_name == 'vgg-16':
model = models.vgg16(pretrained=True).eval().to(device)
mean = torch.Tensor([0.485, 0.456, 0.406])
std = torch.Tensor([0.229, 0.224, 0.225])
if torch.cuda.is_available():
mean = mean.cuda(0)
std = std.cuda(0)
preprocessing = dict(mean=mean, std=std, axis=-3)
fmodel = PyTorchModel(model, bounds=(0, 1), preprocessing=preprocessing)
return fmodel
elif model_name == 'resnet-101':
model = models.resnet101(pretrained=True).eval().to(device)
mean = torch.Tensor([0.485, 0.456, 0.406])
std = torch.Tensor([0.229, 0.224, 0.225])
if torch.cuda.is_available():
mean = mean.cuda(0)
std = std.cuda(0)
preprocessing = dict(mean=mean, std=std, axis=-3)
fmodel = PyTorchModel(model, bounds=(0, 1), preprocessing=preprocessing)
return fmodel
elif model_name == 'densenet-121':
model = models.densenet121(pretrained=True).eval().to(device)
mean = torch.Tensor([0.485, 0.456, 0.406])
std = torch.Tensor([0.229, 0.224, 0.225])
if torch.cuda.is_available():
mean = mean.cuda(0)
std = std.cuda(0)
preprocessing = dict(mean=mean, std=std, axis=-3)
fmodel = PyTorchModel(model, bounds=(0, 1), preprocessing=preprocessing)
return fmodel
def get_label(logit):
_, predict = torch.max(logit, 1)
return predict
def save_results(args,my_intermediates, n):
path = args.output_folder
if not os.path.exists(path):
os.mkdir(path)
if not os.path.exists(path):
os.mkdir(path)
numpy_results = np.full((n * 3, args.max_queries), np.nan)
for i, my_intermediate in enumerate(my_intermediates):
length = len(my_intermediate)
for j in range(length):
numpy_results[3 * i][j] = my_intermediate[j][0]
numpy_results[3 * i + 1][j] = my_intermediate[j][1]
numpy_results[3 * i + 2][j] = my_intermediate[j][2]
pandas_results = pd.DataFrame(numpy_results)
pandas_results.to_csv(os.path.join(path,'results.csv'))
print('save results to:{}'.format(os.path.join(path,'results.csv')))
def read_imagenet_data_specify(args, device):
images = []
labels = []
info = pd.read_csv(args.csv)
selected_image_paths = []
for d_i in range(len(info)):
image_path = info.iloc[d_i]['ImageName']
image = Image.open(os.path.join(args.dataset_path,image_path))
image = image.convert('RGB')
image = image.resize((args.side_length, args.side_length))
image = np.asarray(image, dtype=np.float32)
image = np.transpose(image, (2, 0, 1))
groundtruth = info.iloc[d_i]['Label']
images.append(image)
labels.append(groundtruth)
selected_image_paths.append(image_path)
images = np.stack(images)
labels = np.array(labels)
images = images / 255
images = torch.from_numpy(images).to(device)
labels = torch.from_numpy(labels).to(device).long()
return images, labels, selected_image_paths