-
Notifications
You must be signed in to change notification settings - Fork 0
/
wgan_gp.py
276 lines (219 loc) · 11.6 KB
/
wgan_gp.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
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
import utils
import torch
import time
import os
import pickle
import numpy as np
import torch.nn.parallel
import torch.optim as optim
import torch.utils.data
import torch.utils.data.distributed
import datetime
from torch.autograd import Variable, grad
from torch.utils.data import DataLoader
from visdom import Visdom
from utils.inception_score import inception_score
from sklearn.preprocessing import scale
from dataloader import dataloader
from generator import *
from discriminator import *
class WGAN_GP(object):
def __init__(self, args):
# parameters
self.model_name = args.model_name
self.epoch = args.epoch
self.sample_num = args.sample_num
self.z_dim = args.z_dim
self.input_channels = args.input_channels
self.output_channels = args.output_channels
self.batch_size = args.batch_size
self.save_dir = args.save_dir
self.result_dir = args.result_dir
self.dataset = args.dataset
self.datadir = args.datadir
self.log_dir = args.log_dir
self.generator_arch = args.generator
self.discriminator_arch = args.discriminator
self.nThreads = args.nThreads
self.gpu_mode = args.gpu_mode
self.lambda_ = args.lambda_grad_penalty
self.n_critic = args.n_critic
self.visualize = args.visualize
self.ngpu = args.ngpu
self.env_display = str(args.dataset)
self.vis = Visdom(server=args.visdom_server, port=args.visdom_port)
self.calculate_inception = args.calculate_inception
# networks init
if self.generator_arch == 'infogan':
self.G = INFOGAN_generator(nc=self.input_channels, ngf=self.z_dim)
elif self.generator_arch == 'dcgan':
self.G = DCGAN_generator(self.ngpu, nc=self.input_channels, ngf=self.z_dim)
elif self.generator_arch == 'resnet_6blocks':
self.G = build_generator(self.ngpu, ngf=self.z_dim, which_model_netG='resnet_6blocks')
elif self.generator_arch == 'resnet_9blocks':
self.G = build_generator(self.ngpu, ngf=self.z_dim, which_model_netG='resnet_9blocks')
elif self.generator_arch == 'unet_128':
self.G = build_generator(self.ngpu, ngf=self.z_dim, which_model_netG='unet_128')
elif self.generator_arch == 'unet_256':
self.G = build_generator(self.ngpu, ngf=self.z_dim, which_model_netG='unet_256')
if self.discriminator_arch == 'infogan':
self.D = INFOGAN_discriminator(nc=self.input_channels, ndf=self.z_dim)
elif self.discriminator_arch == 'dcgan':
self.D = DCGAN_discriminator(self.ngpu, nc=self.input_channels, ndf=self.z_dim)
elif self.discriminator_arch == 'basic':
self.D = build_discriminator(self.ngpu, ndf=self.z_dim, which_model_netD='basic')
elif self.discriminator_arch == 'n_layers':
self.D = build_discriminator(self.ngpu, ndf=self.z_dim, which_model_netD='n_layers')
elif self.discriminator_arch == 'pixel':
self.D = build_discriminator(self.ngpu, ndf=self.z_dim, which_model_netD='pixel')
self.G_optimizer = optim.Adam(self.G.parameters(), lr=args.lrG, betas=(args.beta1, args.beta2))
self.D_optimizer = optim.Adam(self.D.parameters(), lr=args.lrD, betas=(args.beta1, args.beta2))
if self.gpu_mode:
self.G.cuda()
self.D.cuda()
print('---------- Networks architecture -------------')
utils.print_network(self.G)
utils.print_network(self.D)
print('-----------------------------------------------')
data = dataloader(self.dataset, self.datadir, self.batch_size, self.nThreads)
self.data_loader = data.load()
print("Data loaded")
#self.z_dim = 512
# fixed noise
if self.gpu_mode:
self.sample_z_ = Variable(torch.rand((self.batch_size, self.input_channels, self.batch_size, self.batch_size)).cuda(), volatile=True)
#self.sample_z_ = Variable(torch.rand((self.batch_size, self.z_dim, 1, 1)).cuda(), volatile=True)
else:
self.sample_z_ = Variable(torch.rand((self.batch_size, self.input_channels, self.batch_size, self.batch_size)), volatile=True)
#self.sample_z_ = Variable(torch.rand((self.batch_size, self.z_dim, 1, 1)), volatile=True)
def train(self):
self.train_hist = {}
self.train_hist['D_loss'] = []
self.train_hist['G_loss'] = []
self.train_hist['per_epoch_time'] = []
self.train_hist['total_time'] = []
if self.gpu_mode:
self.y_real_, self.y_fake_ = Variable(torch.ones((self.batch_size, self.input_channels, self.batch_size, self.batch_size)).cuda()), Variable(torch.zeros((self.batch_size, self.input_channels, self.batch_size, self.batch_size)).cuda())
else:
self.y_real_, self.y_fake_ = Variable(torch.ones((self.batch_size, self.input_channels, self.batch_size, self.batch_size))), Variable(torch.zeros((self.batch_size, self.input_channels, self.batch_size, self.batch_size)))
self.D.train()
print('training start!!')
start_time = time.time()
for epoch in range(self.epoch):
self.G.train()
epoch_start_time = time.time()
for iter, (x_, _) in enumerate(self.data_loader):
if iter == self.data_loader.dataset.__len__() // self.batch_size:
break
z_ = torch.rand((self.batch_size, self.input_channels, self.batch_size, self.batch_size))
#z_ = torch.rand((self.batch_size, self.z_dim, 1, 1))
#z_ = torch.rand(self.batch_size, 3)
if self.gpu_mode:
x_, z_ = Variable(x_.cuda()), Variable(z_.cuda())
else:
x_, z_ = Variable(x_), Variable(z_)
#print(x_.size())
#print(z_.size())
# update D network
self.D_optimizer.zero_grad()
D_real = self.D(x_)
D_real_loss = -torch.mean(D_real)
#print(z_)
G_ = self.G(z_)
#print(G_)
D_fake = self.D(G_)
D_fake_loss = torch.mean(D_fake)
# gradient penalty
if self.gpu_mode:
alpha = torch.rand(x_.size()).cuda()
else:
alpha = torch.rand(x_.size())
# the only reason the imagenet might fail
# is that the miss dimension
x_hat = Variable(alpha * x_.data + (1 - alpha) * G_.data, requires_grad=True)
pred_hat = self.D(x_hat)
if self.gpu_mode:
gradients = grad(outputs=pred_hat, inputs=x_hat, grad_outputs=torch.ones(pred_hat.size()).cuda(),
create_graph=True, retain_graph=True, only_inputs=True)[0]
else:
gradients = grad(outputs=pred_hat, inputs=x_hat, grad_outputs=torch.ones(pred_hat.size()),
create_graph=True, retain_graph=True, only_inputs=True)[0]
gradient_penalty = self.lambda_ * ((gradients.view(gradients.size()[0], -1).norm(2, 1) - 1) ** 2).mean()
D_loss = D_real_loss + D_fake_loss + gradient_penalty
D_loss.backward()
self.D_optimizer.step()
if ((iter+1) % self.n_critic) == 0:
# update G network
self.G_optimizer.zero_grad()
G_ = self.G(z_)
D_fake = self.D(G_)
G_loss = -torch.mean(D_fake)
self.train_hist['G_loss'].append(G_loss.data[0])
G_loss.backward()
self.G_optimizer.step()
self.train_hist['D_loss'].append(D_loss.data[0])
if ((iter + 1) % 100) == 0:
print("Epoch: [%2d] [%4d/%4d] D_loss: %.8f, G_loss: %.8f" %
((epoch + 1), (iter + 1), self.data_loader.dataset.__len__() // self.batch_size, D_loss.data[0], G_loss.data[0]))
self.train_hist['per_epoch_time'].append(time.time() - epoch_start_time)
self.visualize_results((epoch+1))
self.train_hist['total_time'].append(time.time() - start_time)
print("Avg one epoch time: %.2f, total %d epochs time: %.2f" % (np.mean(self.train_hist['per_epoch_time']),
self.epoch, self.train_hist['total_time'][0]))
print("Training finish!... save training results")
self.save()
utils.generate_animation(self.result_dir + '/' + self.dataset + '/' + self.model_name + '/' + self.model_name,
self.epoch)
utils.loss_plot(self.train_hist, os.path.join(self.save_dir, self.dataset, self.model_name), self.model_name)
def visualize_results(self, epoch, fix=True):
self.G.eval()
if not os.path.exists(self.result_dir + '/' + self.dataset + '/' + self.model_name):
os.makedirs(self.result_dir + '/' + self.dataset + '/' + self.model_name)
tot_num_samples = min(self.sample_num, self.batch_size)
image_frame_dim = int(np.floor(np.sqrt(tot_num_samples)))
if fix:
""" fixed noise """
samples = self.G(self.sample_z_)
else:
""" random noise """
if self.gpu_mode:
sample_z_ = Variable(torch.rand((self.batch_size, self.z_dim)).cuda(), volatile=True)
else:
sample_z_ = Variable(torch.rand((self.batch_size, self.z_dim)), volatile=True)
samples = self.G(sample_z_)
copy_samples = samples
if self.gpu_mode:
samples = samples.cpu().data.numpy().transpose(0, 2, 3, 1)
presamples = copy_samples.cpu().data.numpy()
presamples = 2*(presamples - np.max(presamples))/-np.ptp(presamples)-1
else:
samples = samples.data.numpy().transpose(0, 2, 3, 1)
presamples = copy_samples.data.numpy()
presamples = 2*(presamples - np.max(presamples))/-np.ptp(presamples)-1
images = samples[:image_frame_dim * image_frame_dim, :, :, :]
# Calculate inception score
if self.calculate_inception:
score = inception_score(presamples,
cuda=True,
resize=True,
batch_size=self.batch_size)
# display inception mean and std
print("Inception score mean and std are", score)
# save images and display if set
utils.save_images(self.vis, images, [image_frame_dim, image_frame_dim],
self.result_dir + '/' + self.dataset + '/' + self.model_name + '/'
+ self.model_name + '_epoch%03d' % epoch + '.png',
env=self.env_display,
visualize=self.visualize)
def save(self):
save_dir = os.path.join(self.save_dir, self.dataset, self.model_name)
if not os.path.exists(save_dir):
os.makedirs(save_dir)
torch.save(self.G.state_dict(), os.path.join(save_dir, self.model_name + '_G.pkl'))
torch.save(self.D.state_dict(), os.path.join(save_dir, self.model_name + '_D.pkl'))
with open(os.path.join(save_dir, self.model_name + '_history.pkl'), 'wb') as f:
pickle.dump(self.train_hist, f)
def load(self):
save_dir = os.path.join(self.save_dir, self.dataset, self.model_name)
self.G.load_state_dict(torch.load(os.path.join(save_dir, self.model_name + '_G.pkl')))
self.D.load_state_dict(torch.load(os.path.join(save_dir, self.model_name + '_D.pkl')))