Repository navigation
Expand file tree
/
Copy pathmain.py
More file actions
312 lines (252 loc) · 11.2 KB
/
Copy pathmain.py
File metadata and controls
312 lines (252 loc) · 11.2 KB
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
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
# Import system libraries
import argparse
import math
import os
import random
import sys
import time
# import 3rd party libraries
import csv
import matplotlib.pyplot as plt
import skimage
import torch
import torchvision
from PIL import Image
from torch import nn , optim
from torch.autograd import Variable
from torchvision import datasets, transforms
# Import user-defined libraries
from dataset import DatasetFromFolder
from network import EncoderNet, DecoderNet
irange = range
parser = argparse.ArgumentParser(description='Image Compression')
parser.add_argument('--batchSize', type=int, default=64, help='training batch size')
parser.add_argument('--nEpochs', type=int, default=2, help='number of epochs to train for')
parser.add_argument('--lr', type=float, default=0.001, help='Learning Rate. Default=0.001')
parser.add_argument('--beta', type=float, default=0.99, help='beta1 for adam. default=0.99')
parser.add_argument('--threads', type=int, default=4, help='number of threads for data loader to use')
parser.add_argument('--encoder_net', type=str, default='', help='Path to pre-trained encoder net. Default=3')
parser.add_argument('--decoder_net', type=str, default='', help='path to pre-trained deocder net. Default=3')
parser.add_argument('--outf', default='.', help='folder to output images and model checkpoints')
parser.add_argument('--test_path', default='./Dataset/test', help='path of test images')
parser.add_argument('--channels', type=int, default=3, help='number of channels in an image. Default=3')
parser.add_argument('--dataset', type=str, default='STL10', help='dataset to be used for training and validation. Default=STL10')
parser.add_argument('--data_path', type=str, default='./Dataset/CLIC', help='path to images. Default=CLIC')
parser.add_argument('--image_size', type=int, default=90, help='path to images. Default=90')
parser.add_argument('--loss_function', type=int, default=0, help='Loss function. Default=0')
parser.add_argument('--use_GPU', type=int, default=-1, help='0 for GPU, 1 for CPU . Default=AUTO')
parser.add_argument('--mode', type=str, default='both', help='train / test / both . Default=both')
opt = parser.parse_args()
print (opt)
CUDA = torch.cuda.is_available()
LOG_INTERVAL = 5
if opt.use_GPU == 1:
CUDA = False
def save_image(tensor, filename, nrow=8, padding=2,
normalize=False, range=None, scale_each=False, pad_value=0):
"""Save a given Tensor into an image file.
Args:
tensor (Tensor or list): Image to be saved. If given a mini-batch tensor,
saves the tensor as a grid of images by calling ``make_grid``.
**kwargs: Other arguments are documented in ``make_grid``.
"""
torchvision.utils.save_image(tensor,filename)
def img_transform_train(crop_size):
if opt.dataset.upper() == 'STL10':
if opt.channels == 1 :
return transforms.Compose([
transforms.Resize(crop_size),
transforms.Grayscale(num_output_channels=1),
transforms.ToTensor(),
])
else :
return transforms.Compose([
transforms.Resize(crop_size),
transforms.ToTensor(),
])
else :
return transforms.Compose([
transforms.RandomApply([
transforms.RandomHorizontalFlip(p=0.8),
transforms.RandomVerticalFlip(),
transforms.RandomRotation(degrees= random.randint(10,350))
]),
transforms.Resize(crop_size),
transforms.ToTensor(),
])
def img_transform_test(crop_size):
return transforms.Compose([
transforms.Resize(crop_size),
transforms.ToTensor(),
])
"""### Loss function"""
def loss_function(final_img,residual_img,upscaled_img,com_img,orig_img):
if opt.loss_function == 0:
com_loss = nn.MSELoss()(orig_img, final_img)
rec_loss = nn.MSELoss()(residual_img,orig_img-upscaled_img)
return com_loss + rec_loss
elif opt.loss_function == 1:
com_loss = nn.MSELoss()(orig_img, final_img)
rec_loss = torch.sum(torch.std(residual_img))
return com_loss + rec_loss
"""### define Train and Test methods"""
def train(epoch,model1,model2,train_loader):
optimizer1 = optim.Adam(model1.parameters(), lr=1e-3)
optimizer2 = optim.Adam(model2.parameters(), lr=1e-3)
model1.train()
model2.train()
train_loss = 0
for batch_idx, data in enumerate(train_loader):
if opt.dataset.upper() == 'STL10':
data , _ = data
data = Variable(data)
optimizer1.zero_grad()
optimizer2.zero_grad()
if CUDA:
data = data.cuda()
com_img = model1(data)
final, residual_img, upscaled_image = model2(com_img)
else :
data = data.cpu()
com_img = model1(data)
final, residual_img, upscaled_image = model2(com_img)
loss = loss_function(final, residual_img, upscaled_image, com_img, data)
loss.backward()
train_loss += loss.item()
optimizer1.step()
optimizer2.step()
if batch_idx % LOG_INTERVAL == 0:
print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
epoch, batch_idx * len(data), len(train_loader.dataset),
100. * batch_idx / len(train_loader),
loss.item() / len(data)))
print('====> Epoch: {} Average loss: {:.4f}'.format(
epoch, train_loss / len(train_loader.dataset)))
return train_loss / len(train_loader.dataset)
def validation(epoch,model1,model2, val_loader):
model1.eval()
model2.eval()
val_loss = 0
total_psnr = 0
torch.no_grad()
for i, data in enumerate(val_loader):
if opt.dataset.upper() == 'STL10':
data , _ = data
data = Variable(data)
if CUDA:
data = data.cuda()
com_img = model1(data)
final, residual_img, upscaled_image = model2(com_img)
else :
data = data.cpu()
com_img = model1(data)
final, residual_img, upscaled_image = model2(com_img)
batch_loss = loss_function(final, residual_img, upscaled_image, com_img, data).item()
val_loss += batch_loss
psnr = 10 * math.log10(1 / batch_loss)
total_psnr += psnr
val_loss /= len(val_loader.dataset)
print('====> val set loss: {:.4f}'.format(val_loss))
print('====> Avg. PSNR : {:.4f}'.format(total_psnr/len(val_loader.dataset)))
return val_loss
def test(model1,model2,test_loader):
t1 = time.time()
model1.load_state_dict(torch.load('./Encoder_net.pth'))
model2.load_state_dict(torch.load('./Decoder_net.pth'))
model1.eval()
model2.eval()
test_loss = 0
torch.no_grad()
for i, data in enumerate(test_loader):
data = Variable(data)
if CUDA:
data = data.cuda()
com_img = model1(data)
final, residual_img, upscaled_image = model2(com_img)
else:
data = data.cpu()
com_img = model1(data)
final, residual_img, upscaled_image = model2(com_img)
test_loss += loss_function(final, residual_img, upscaled_image, com_img, data).item()
n = min(data.size(0), 6)
print("saving the image "+str(i))
comparison = torch.cat([data[:i],
final[:i]])
comparison = comparison.cpu()
save_image(com_img[0].data,
'compressed_' + str(i) +'.png', nrow=n)
save_image(final[0].data,
'final_' + str(i) +'.png', nrow=n)
save_image(data[0].data,
'original_' + str(i) +'.png', nrow=n)
test_loss /= len(test_loader.dataset)
print('====> Test set loss: {:.4f}'.format(test_loss))
print("Avg. Time required for inference :"+str((time.time()-t1)/len(test_loader.dataset)))
def main():
"""### Parameters"""
CHANNELS = opt.channels
HEIGHT = opt.image_size
EPOCHS = opt.nEpochs
"""### Load Dataset"""
if opt.dataset.upper() == 'STL10':
trainset = datasets.STL10(root='./data', split='train',download=True, transform=img_transform_train((opt.image_size,opt.image_size)))
val_set = datasets.STL10(root='./data', split='test',download=True, transform=img_transform_train((opt.image_size,opt.image_size)))
elif opt.dataset.upper() == 'FOLDER':
trainset = DatasetFromFolder(opt.data_path+'train/', input_transform=img_transform_train((opt.image_size,opt.image_size)),channels = opt.channels)
val_set = DatasetFromFolder(opt.data_path+'valid/', input_transform=img_transform_train((opt.image_size,opt.image_size)),channels = opt.channels)
train_loader = torch.utils.data.DataLoader(trainset, batch_size=opt.batchSize,
shuffle=True, num_workers=opt.threads)
val_loader = torch.utils.data.DataLoader(val_set, batch_size=opt.batchSize,
shuffle=False, num_workers=opt.threads)
test_set = DatasetFromFolder(opt.test_path,
input_transform=img_transform_test((opt.image_size,opt.image_size)),channels = opt.channels)
test_data_loader = torch.utils.data.DataLoader(dataset=test_set, num_workers=opt.threads, shuffle=False)
info = {}
info['channels'] = CHANNELS
info['size'] = HEIGHT
if CUDA:
model1 = EncoderNet(info).cuda()
model2 = DecoderNet(info).cuda()
else :
model1 = EncoderNet(info).cpu()
model2 = DecoderNet(info).cpu()
print("GPU available : "+str(CUDA))
if opt.mode.upper() == 'TEST':
print(" Mode selected : test")
print("run model for Images in test folder.")
test(model1,model2,test_data_loader)
sys.exit()
"""### Program Execution"""
tr_loss = []
vl_loss = []
for epoch in range(1, EPOCHS+1):
t1 = time.time()
print("Epoch : "+str(epoch))
t_loss = train(epoch,model1,model2,train_loader)
v_loss = validation(epoch,model1,model2,val_loader)
tr_loss.append(t_loss)
vl_loss.append(v_loss)
print("Time required for "+str(epoch)+"th epoch: "+str(time.time() - t1))
if epoch % 50 == 0:
torch.save(model1.state_dict(), '%s/Encoder_epoch_%d.pth' % (opt.outf, epoch))
torch.save(model2.state_dict(), '%s/Decoder_epoch_%d.pth' % (opt.outf, epoch))
print("Plot and save train and validation loss curves")
# Plot train, test loss curves
plt.plot(range(EPOCHS),tr_loss , 'r--',label='Training Loss')
plt.plot(range(EPOCHS), vl_loss, 'b--',label='Validation Loss')
plt.title('Loss vs Epochs')
plt.legend()
plt.xlabel('Epochs')
plt.ylabel('loss')
plt.savefig('loss_'+str(time.time())+'.png')
with open('loss_values.csv', "w") as output:
writer = csv.writer(output, lineterminator='\n')
writer.writerows([tr_loss,vl_loss])
print("save the models")
torch.save(model1.state_dict(), './Encoder_net.pth')
torch.save(model2.state_dict(), './Decoder_net.pth')
if opt.mode.upper() == 'BOTH':
print("run model for Images in test folder.")
test(model1,model2,test_data_loader)
if __name__ == '__main__':
main()