-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathmodel_params.py
141 lines (112 loc) · 6.05 KB
/
model_params.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
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
'''
In this file one can set the hyperparameters for training, which model architectures to use
as well as include or exclude out-distribution datasets that one wishes to test on
Nearly all scripts reference this file
'''
import utils.models as models
import utils.dataloaders as dl
import utils.resnet_orig as resnet
class MNIST_params():
def __init__(self, augm_flag=True, batch_size=128):
self.data_name = 'MNIST'
self.dim = 784
self.base_model = models.LeNetMadry()
self.train_loader = dl.MNIST(train=True, batch_size=batch_size, augm_flag=augm_flag)
self.cali_loader = dl.MNIST(train=True, batch_size=batch_size, augm_flag=False)
self.test_loader = dl.MNIST(train=False)
self.loaders = [('FMNIST', dl.FMNIST(train=False)),
('EMNIST', dl.EMNIST(train=False)),
('GrayCIFAR10', dl.GrayCIFAR10(train=False)),
# ('TinyImages', dl.TinyImages(self.data_name, batch_size=batch_size, train=False)),
('Noise', dl.Noise(dataset=self.data_name, batch_size=batch_size)),
('UniformNoise', dl.UniformNoise(self.data_name, batch_size=batch_size))]
self.tinyimage_loader = dl.TinyImages(self.data_name, batch_size=100)
self.data_used = 60000
self.epsilon = 0.3
self.lr = 1e-3
self.classes = 10
class FMNIST_params():
def __init__(self, augm_flag=True, batch_size=128):
self.data_name = 'FMNIST'
self.base_model = resnet.ResNet18(num_of_channels=1)
self.train_loader = dl.FMNIST(train=True, batch_size=batch_size, augm_flag=augm_flag)
self.cali_loader = dl.FMNIST(train=True, batch_size=batch_size, augm_flag=False)
self.test_loader = dl.FMNIST(train=False, augm_flag=False)
self.dim = 784
self.loaders = [('MNIST', dl.MNIST(train=False)),
('EMNIST', dl.EMNIST(train=False)),
('GrayCIFAR10', dl.GrayCIFAR10(train=False)),
# ('TinyImages', dl.TinyImages(self.data_name, batch_size=batch_size, train=False)),
('Noise', dl.Noise(dataset=self.data_name, batch_size=batch_size)),
('UniformNoise', dl.UniformNoise(dataset=self.data_name, batch_size=batch_size))]
self.tinyimage_loader = dl.TinyImages(self.data_name, batch_size=100)
self.data_used = 60000
self.epsilon = 0.3
self.lr = 0.1
self.classes = 10
class SVHN_params():
def __init__(self, augm_flag=True, batch_size=128):
self.data_name = 'SVHN'
self.base_model = resnet.ResNet18()
self.train_loader = dl.SVHN(train=True, batch_size=batch_size, augm_flag=augm_flag)
self.cali_loader = dl.SVHN(train=True, batch_size=batch_size, augm_flag=False)
self.test_loader = dl.SVHN(train=False)
self.dim = 3072
self.loaders = [('CIFAR10', dl.CIFAR10(train=False)),
('CIFAR100', dl.CIFAR100(train=False)),
('LSUN_CR', dl.LSUN_CR(train=False)),
('Imagenet-',dl.ImageNetMinusCifar10(train=False)),
# ('TinyImages', dl.TinyImages(self.data_name, batch_size=batch_size, train=False)),
('Noise', dl.Noise(dataset='SVHN', batch_size=batch_size)),
('UniformNoise', dl.UniformNoise(dataset=self.data_name, batch_size=batch_size))]
self.tinyimage_loader = dl.TinyImages(self.data_name, batch_size=100)
self.data_used = 50000
self.epsilon = 0.3
self.lr = 0.1
self.classes = 10
class CIFAR10_params():
def __init__(self, augm_flag=True, batch_size=128):
self.data_name = 'CIFAR10'
self.base_model = resnet.ResNet18()
self.train_loader = dl.CIFAR10(train=True, batch_size=batch_size, augm_flag=augm_flag)
self.cali_loader = dl.CIFAR10(train=True, batch_size=batch_size, augm_flag=False)
self.test_loader = dl.CIFAR10(train=False)
self.dim = 3072
self.loaders = [('SVHN', dl.SVHN(train=False)),
('CIFAR100', dl.CIFAR100(train=False)),
('LSUN_CR', dl.LSUN_CR(train=False)),
('Imagenet-',dl.ImageNetMinusCifar10(train=False)),
# ('TinyImages', dl.TinyImages(self.data_name, batch_size=batch_size, train=False)),
('Noise', dl.Noise(dataset='CIFAR10', batch_size=batch_size)),
('UniformNoise', dl.UniformNoise(dataset=self.data_name, batch_size=batch_size))]
self.tinyimage_loader = dl.TinyImages(self.data_name, batch_size=100)
self.data_used = 50000
self.epsilon = 0.3
self.lr = 0.1
self.classes = 10
class CIFAR100_params():
def __init__(self, augm_flag=True, batch_size=128):
self.data_name = 'CIFAR100'
self.base_model = resnet.ResNet18(num_classes=100)
self.train_loader = dl.CIFAR100(train=True, batch_size=batch_size, augm_flag=augm_flag)
self.cali_loader = dl.CIFAR100(train=True, batch_size=batch_size, augm_flag=False)
self.test_loader = dl.CIFAR100(train=False)
self.dim = 3072
self.loaders = [('SVHN', dl.SVHN(train=False)),
('CIFAR10', dl.CIFAR10(train=False)),
('LSUN_CR', dl.LSUN_CR(train=False)),
('Imagenet-',dl.ImageNetMinusCifar10(train=False)),
# ('TinyImages', dl.TinyImages(self.data_name, batch_size=batch_size, train=False)),
('Noise', dl.Noise(dataset='CIFAR100', batch_size=batch_size)),
('UniformNoise', dl.UniformNoise(dataset=self.data_name, batch_size=batch_size))]
self.tinyimage_loader = dl.TinyImages(self.data_name, batch_size=100)
self.data_used = 50000
self.epsilon = 0.3
self.lr = 0.1
self.classes = 100
params_dict = {'MNIST': MNIST_params,
'FMNIST': FMNIST_params,
'SVHN': SVHN_params,
'CIFAR10': CIFAR10_params,
'CIFAR100': CIFAR100_params,
}