-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathold_gan_mnist.py
151 lines (124 loc) · 5.98 KB
/
old_gan_mnist.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
# pylint: disable=E0401
import torchvision.datasets as ds
import torchvision.transforms as transforms
from torch.autograd import Variable
import torch
import torch.nn as nn
import torch.optim as optim
import torch.nn.functional as F
import numpy as np
import torchvision.utils as vutils
transform = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,))])
batch_size = 200
gen_noise_size = 100
epochs = 30
noise_deviaiton = 0.2
k_d, k_g = 1,1
image_size = 28
fixed_z = Variable(torch.randn(batch_size, gen_noise_size)).cuda()
sw = 4
num_weights = 64
nc = 1
model_out_folder = "./checkpoints"
pic_out_folder = "./sample_images"
class Generator(nn.Module):
def __init__(self):
super(Generator, self).__init__()
self.fc1 = nn.Linear(gen_noise_size, 4*num_weights*sw*sw)
self.norm1 = nn.BatchNorm2d(4*num_weights)
self.dc1 = nn.ConvTranspose2d(num_weights*4, num_weights*2, 4, stride=1, padding=0)
self.norm2 = nn.BatchNorm2d(2*num_weights)
self.dc2 = nn.ConvTranspose2d(num_weights*2, num_weights, 4, stride=2, padding=1)
self.norm3 = nn.BatchNorm2d(num_weights)
self.dc3 = nn.ConvTranspose2d(num_weights, nc, 4, stride=2, padding=1)
def forward(self, x):
x = self.norm1(F.relu(self.fc1(x)).view(batch_size, num_weights*4, sw, sw))
x = self.norm2(F.relu(self.dc1(x)))
x = self.norm3(F.relu(self.dc2(x)))
x = self.dc3(x)
return F.tanh(x)
class Discriminator(nn.Module):
def __init__(self):
super(Discriminator, self).__init__()
self.conv1 = nn.Conv2d(nc, num_weights, kernel_size=4, stride=2, padding=1)
self.norm1 = nn.BatchNorm2d(num_weights)
self.conv2 = nn.Conv2d(num_weights, num_weights*2, kernel_size=4, stride=2, padding=1)
self.norm2 = nn.BatchNorm2d(2*num_weights)
self.conv3 = nn.Conv2d(num_weights*2, num_weights*4, kernel_size=4, stride=1, padding=0)
self.norm3 = nn.BatchNorm2d(4*num_weights)
self.fc1 = nn.Linear(4*num_weights*sw*sw, 1)
def forward(self, x):
x = self.norm1(F.leaky_relu(self.conv1(x)))
x = self.norm2(F.leaky_relu(self.conv2(x)))
x = self.norm3(F.leaky_relu(self.conv3(x)))
x = self.fc1(x.view(batch_size, -1))
return F.sigmoid(x)
generator = Generator()
generator = generator.cuda()
discriminator = Discriminator()
discriminator = discriminator.cuda()
criterion = nn.BCELoss()
gen_optimizer = optim.Adam(generator.parameters(), lr=0.001)
dis_optimizer = optim.Adam(discriminator.parameters(), lr=0.001)
trainset = ds.MNIST("./mnist", download=True, train=True, transform=transform)
testset = ds.MNIST("./mnist", download=True, train=False, transform=transform)
# trainset = torch.index_select(train_mnist.train_data, 0, torch.LongTensor(np.where(train_mnist.train_labels.numpy() == 2)[0])).type(torch.FloatTensor)
# testset = torch.index_select(test_mnist.test_data, 0, torch.LongTensor(np.where(test_mnist.test_labels.numpy() == 2)[0])).type(torch.FloatTensor)
train_feeder = torch.utils.data.DataLoader(
trainset, batch_size=batch_size, shuffle=True, num_workers=2, drop_last=True)
test_feeder = torch.utils.data.DataLoader(
testset, batch_size=batch_size, shuffle=False, num_workers=2, drop_last=True)
for epoch in range(epochs): # loop over the dataset multiple times
running_dis_loss, running_gen_loss = 0.0, 0.0
for i, data in enumerate(train_feeder, 0):
inputs, labels = data
real_samples = Variable(inputs.cuda())
# labels = Variable(labels.cuda())
# train discriminator
z = Variable(torch.randn((batch_size, gen_noise_size))).cuda()
fake_samples = generator(z)
# zero the parameter gradients
dis_optimizer.zero_grad()
fake_labels = discriminator(fake_samples.view(batch_size,1,28,28).detach())
#train on generated samples
fake_loss = criterion(fake_labels, Variable(torch.FloatTensor(batch_size).fill_(0)).cuda())
fake_loss.backward()
real_labels = discriminator(real_samples)
#train on real samples
real_loss = criterion(real_labels, Variable(torch.FloatTensor(batch_size).fill_(1)).cuda())
real_loss.backward()
dis_optimizer.step()
# train generator
gen_optimizer.zero_grad()
dis_labels = discriminator(fake_samples)
gen_loss = criterion(dis_labels, Variable(torch.FloatTensor(batch_size).fill_(1)).cuda())
gen_loss.backward()
gen_optimizer.step()
# print loss statistics
running_dis_loss += real_loss.data.mean() + fake_loss.data.mean()
running_gen_loss += gen_loss.data.mean()
if i % 20 == 19: # print every 2000 mini-batches
print('[%d, %5d] discrimenator loss: %.3f, generator loss: %.3f' %
(epoch + 1, i + 1, running_dis_loss / 2000, running_gen_loss / 2000))
running_dis_loss, running_gen_loss = 0.0, 0.0
if i % 90 == 89:
vutils.save_image(inputs[0],
'%s/real_samples.png' % pic_out_folder,
normalize=True)
fake = generator(fixed_z)
vutils.save_image(fake.data,
'%s/epoch_%d.png' % (pic_out_folder, epoch),
normalize=True)
torch.save(generator.state_dict(), '%s/netG_epoch_%d.pth' % (model_out_folder, epoch))
torch.save(discriminator.state_dict(), '%s/netD_epoch_%d.pth' % (model_out_folder, epoch))
print('Finished Training')
# correct = 0
# total = 0
# for data in test_feeder:
# images, labels = data
# outputs = generator(Variable(images.cuda()))
# _, predicted = torch.max(outputs.data, 1)
# total += labels.size(0)
# correct += (predicted == labels.cuda()).sum()
# print('Accuracy of the network on the 10000 test images: %d %%' % (
# 100 * correct / total))