-
Notifications
You must be signed in to change notification settings - Fork 0
/
cmnist_leftovers.py
168 lines (139 loc) · 5.39 KB
/
cmnist_leftovers.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
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
#Experiment Objectives:
#1. Generate Attack from model j and test accuracies, L2 Distances and Reconstructions on model i (!=j)
import torch
import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
from tqdm import tqdm
import os
import matplotlib.pyplot as plt
from advertorch.attacks import LinfPGDAttack, GradientSignAttack, CarliniWagnerL2Attack, LinfBasicIterativeAttack
# import seaborn as sns
import sys
sys.path.insert(0,'/raid/sdas_ma/Adversarial_CapsNet_Pytorch/')
from model.net import *
from model.cnn_net import *
from utils.training import *
from data.data import *
base_path = '/raid/sdas_ma/Adversarial_CapsNet_Pytorch/'
model_path = "/raid/sdas_ma/Adversarial_CapsNet_Pytorch/weights/"#os.path.join(os.getcwd(), "weights")
Caps_args = {
'DATASET_NAME':'mnist',
'num_classes':10,
'USE_CUDA': True if torch.cuda.is_available() else False,
'BATCH_SIZE': 512,
##For Decoder
'num_features':160,
'LReLU_negative_slope':0.1,
'input_height':28,
'input_width':28,
'input_channel':1,
}
CNN_args = {
'DATASET_NAME':'mnist',
'num_classes':10,
'USE_CUDA': True if torch.cuda.is_available() else False,
'BATCH_SIZE': 256,
#For Decoder
'num_features':160,
'LReLU_negative_slope':0.1,
'input_height':28,
'input_width':28,
'input_channel':1,
'type':'plusCR',
}
class Caps_Config:
def __init__(self, dataset='mnist'):
# CNN (cnn)
self.cnn_in_channels = 1
self.cnn_out_channels = 12
self.cnn_kernel_size = 15
# Primary Capsule (pc)
self.pc_num_capsules = 1
self.pc_in_channels = 12
self.pc_out_channels = 16
self.pc_kernel_size = 8
self.pc_num_routes = 7 * 7
# Digit Capsule 1 (dc)
self.dc_num_capsules = 49
self.dc_num_routes = 7 * 7
self.dc_in_channels = 16
self.dc_out_channels = 16 #1
# Digit Capsule 2 (dc)
self.dc_2_num_capsules = 10
self.dc_2_num_routes = 7 * 7
self.dc_2_in_channels = 16 #1
self.dc_2_out_channels = 16
# Decoder
self.input_width = 28
self.input_height = 28
class CNN_Config:
def __init__(self, dataset='mnist'):
# CONV1
self.conv1_in = 1
self.conv1_out = 12
self.conv1_kernel_size = 15
# CONV2
self.conv2_in = 12
self.conv2_out = 16
self.conv2_kernel_size = 8
# FC1
self.fc1_in = 7 * 7 * 16
self.fc1_out = 784
# FC1
self.fc2_in = 784
self.fc2_out = 160
torch.manual_seed(1)
################Loading Dataset####################
cmnistlist = ["contrast","elastic_transform" ,"gaussian_blur" ,"inverse" ,"line" ,"pixelate" ,"saturate" ,"speckle_noise", "defocus_blur" ,"frost" ,"gaussian_noise" ,"jpeg_compression" ,"pessimal_noise" ,"quantize" ,"snow" ,"zoom_blur"]
cmnistpath = os.path.join(base_path,"data/mnist_c_leftovers/")
def test_n_l2(model, model_name):
net.eval()
Und_l2 = {key:torch.tensor([],dtype=torch.bool) for key in cmnistlist}
l2_distances_all = {key:torch.tensor([],dtype=torch.int16) for key in cmnistlist}
for corruption in cmnistlist:
datapath = os.path.join(cmnistpath, corruption)
unnorm_data = torch.tensor(np.load(os.path.join(datapath, "test_images.npy"))/255).permute((0,3,1,2))
labels = torch.tensor(np.load(os.path.join(datapath, "test_labels.npy"))).reshape(-1)
if(labels.min()==1):
labels -= 1
unnorm_data, labels = unnorm_data.cuda().type(torch.cuda.FloatTensor), labels.cuda()
#normalizing
data = (unnorm_data - 0.1307)/0.3081
with torch.no_grad():
output, reconstructions, max_length_indices = net(data)
l2_distances = ((reconstructions.view(unnorm_data.size(0),-1)-unnorm_data.view(unnorm_data.size(0), -1))**2).sum(1).squeeze().detach()
Und_l2[corruption] = (max_length_indices != labels).detach().cpu()
l2_distances_all[corruption] = l2_distances.detach().cpu()
Und_l2[corruption] = Und_l2[corruption].numpy()
print(model_name," : " ,corruption," : ",np.sum(Und_l2[corruption])/100," // ",np.count_nonzero(l2_distances_all[corruption][Und_l2[corruption]]<45)/100)
np.save(os.path.join(base_path , "results", str("cmnist_leftovers"+model_name+"Undl2.npy")), Und_l2)
np.save(os.path.join(base_path , "results", str("cmnist_leftovers"+model_name+"l2_dist.npy")), l2_distances_all)
def capsnet():
config = Caps_Config()
net = CapsNet(Caps_args, config)
# capsule_net = torch.nn.DataParallel(capsule_net)
if Caps_args['USE_CUDA']:
net = net.cuda()
net.load_state_dict(torch.load(os.path.join(model_path, 'CapsNet_mnist.pth'), map_location='cpu'))
return net
def CNN(model_type):
CNN_args['type'] = model_type
config = CNN_Config()
net = CNNnet(CNN_args, config)
net.load_state_dict(torch.load(os.path.join(model_path, 'CNN'+model_type+'_mnist.pth'), map_location='cpu'))
if CNN_args['USE_CUDA']:
net = net.cuda()
return net
def load_model(model_name):
if(model_name=="capsnet"):
net = capsnet()
return net
else:
net = CNN(model_name)
return net
model_name_list = ["capsnet", "plusCR", "plusR"]
for model_name in model_name_list:
net = load_model(model_name)
test_n_l2(net, model_name)