-
Notifications
You must be signed in to change notification settings - Fork 46
/
main.py
172 lines (131 loc) · 6.14 KB
/
main.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
import os
import time
from collections import OrderedDict
import logging
import argparse
import numpy as np
import random
import torch
import torch.nn as nn
import torch.optim
import torch.utils.data
import torch.backends.cudnn
import torchvision.utils
from torch.autograd import Variable
import torchvision.utils as vutils
from torchvision.utils import save_image
torch.backends.cudnn.benchmark = True
import matplotlib.pyplot as plt
from dataset import get_loader
import imageio
from model import *
def main():
parser = argparse.ArgumentParser()
# model config
parser.add_argument('--image_size', type=int, default=32)
parser.add_argument('--num_class', type=int, default=10)
parser.add_argument('--latent_dim', type=int, default=100)
# run config
parser.add_argument('--outdir', type=str, default='./result/')
parser.add_argument('--seed', type=int, default=17)
parser.add_argument('--num_workers', type=int, default=2)
parser.add_argument('--ndata', type=str, required=True)
# optim config
parser.add_argument('--epochs', type=int, default=50)
parser.add_argument('--batch_size', type=int, default=32)
parser.add_argument('--lr', type=float, default=0.0002)
args = parser.parse_args()
os.environ['CUDA_VISIBLE_DEVICES'] = '0'
args.device = torch.device('cuda:0') if torch.cuda.is_available() else 'cpu'
args.channel = 3 if args.ndata=='cifar10' else 1
seed = args.seed
torch.manual_seed(seed)
random.seed(seed)
outdir = args.outdir
if not os.path.exists(outdir):
os.makedirs(outdir)
os.makedirs(outdir+'generator_state/')
os.makedirs(outdir+'discriminator_state/')
os.makedirs(outdir+'Image/')
dataloader = get_loader(args.batch_size, args.num_workers, args.image_size, args.ndata)
generator = Generator(args).to(args.device)
gen_optimizer = torch.optim.Adam(generator.parameters(), lr=args.lr, betas=(0.5, 0.999))
discriminator = Discriminator(args).to(args.device)
discriminator.apply(init_weights)
d_optimizer = torch.optim.Adam(discriminator.parameters(), lr=args.lr, betas=(0.5, 0.999))
loss = torch.nn.BCELoss()
label_type = torch.LongTensor
img_type = torch.FloatTensor
# Sample noise
fix_z = Variable(torch.FloatTensor(np.random.normal(0, 1, (args.num_class ** 2, args.latent_dim)))).to('cuda:0')
# Get labels ranging from 0 to n_classes for n rows
fix_labels = np.array([num for _ in range(args.num_class) for num in range(args.num_class)])
fix_labels = Variable(torch.LongTensor(fix_labels)).to('cuda:0')
G_losses, D_losses = train(args,generator,discriminator,dataloader,loss, img_type, label_type,gen_optimizer,d_optimizer,fix_labels,fix_z)
# Plotting the loss graph
plt.plot(G_losses, label='Generator')
plt.plot(D_losses, label='Discriminator')
plt.legend()
plt.savefig(args.outdir+"plot.png")
plt.show()
# Making the GIF
image = []
for i in range(1,args.epochs+1):
image.append(imageio.imread(args.outdir+'Image/'+str(i)+'.png'))
imageio.mimsave(args.ndata+'.gif', image, fps=5)
def sample_image(args, z, labels, batches_done, generator):
"""Saves a grid of generated digits ranging from 0 to n_classes"""
gen_imgs = generator(z, labels)
save_image(gen_imgs.data, args.outdir + "Image/%d.png" % batches_done, nrow=args.num_class, normalize=True)
def train(args, generator, discriminator, dataloader, loss, img_type, label_type, gen_optimizer, d_optimizer, fix_label, fix_noise):
generator.train()
discriminator.train()
G_losses = []
D_losses = []
for epoch in range(1, args.epochs + 1):
G_loss = 0.
D_loss = 0.
start_time = time.time()
for i, data in enumerate(dataloader):
(imgs, labels) = data
batch_size = imgs.shape[0]
# print(batch_size)
imgs = Variable(imgs.type(img_type)).to(args.device)
labels = Variable(labels.type(label_type)).to(args.device)
# Creating real and fake label for calculation of loss
r_label = Variable(img_type(batch_size, 1).fill_(0.9)).to(args.device)
f_label = Variable(img_type(batch_size, 1).fill_(0.0)).to(args.device)
# Training Generator
gen_optimizer.zero_grad()
noise = Variable(img_type(np.random.normal(0, 1, (batch_size, args.latent_dim)))).to(args.device)
rand_label = Variable(label_type(np.random.randint(0, args.num_class, batch_size))).to(args.device)
dis = discriminator(generator(noise, rand_label), rand_label)
# print(type(dis),' ',type(r_label))
g_loss = loss(dis, r_label)
g_loss.backward()
gen_optimizer.step()
# Training Discriminator
d_optimizer.zero_grad()
noise = Variable(img_type(np.random.normal(0, 1, (batch_size, args.latent_dim)))).to(args.device)
rand_label = Variable(label_type(np.random.randint(0, args.num_class, batch_size))).to(args.device)
d_real = discriminator(imgs, labels)
loss_real = loss(d_real, r_label)
d_fake = discriminator(generator(noise, rand_label).detach(), rand_label)
# print(d_fake.shape," ",f_label.shape)
loss_fake = loss(d_fake, f_label)
d_loss = loss_fake + loss_real
d_loss.backward()
d_optimizer.step()
G_loss += g_loss.item()
D_loss += d_loss.item()
print('Epoch {} || G_loss: {} || D_loss: {} || Time elapsed: {}'.format(epoch, G_loss / (i), D_loss / (i),
time.time() - start_time))
G_losses.append(G_loss / (i))
D_losses.append(D_loss / (i))
sample_image(args, fix_noise, fix_label, epoch, generator)
# Checkpoint
torch.save(generator.state_dict(), args.outdir + 'generator_state/generator_{}_.pth'.format(epoch))
torch.save(discriminator.state_dict(), args.outdir + 'discriminator_state/discriminator_{}_.pth'.format(epoch))
return G_losses, D_losses
if __name__ == '__main__':
main()