-
Notifications
You must be signed in to change notification settings - Fork 7
/
main.py
197 lines (152 loc) · 8.9 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
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
#! /usr/bin/python
import os, time, pickle, random, time
from datetime import datetime
import numpy as np
from time import localtime, strftime
import logging, scipy
import tensorflow as tf
import tensorlayer as tl
from model import IRCNN
from utils import *
from config import config, log_config
os.environ["CUDA_VISIBLE_DEVICES"]= '1'
###====================== HYPER-PARAMETERS ===========================###
## Adam
batch_size = config.TRAIN.batch_size
lr_init = config.TRAIN.lr_init
beta1 = config.TRAIN.beta1
## learning IRCNN
n_epoch = config.TRAIN.n_epoch
lr_decay = config.TRAIN.lr_decay
decay_every = config.TRAIN.decay_every
ni = int(np.sqrt(batch_size))
def train():
logging.basicConfig(level=logging.DEBUG,
filename='train_info.log',
filemode='w',
format='%(asctime)s - %(pathname)s[line:%(lineno)d] - %(levelname)s: %(message)s'
)
# create folders to save tested result images
save_dir_ircnn = "samples/{}_ircnn".format(tl.global_flag['mode'])
tl.files.exists_or_mkdir(save_dir_ircnn)
save_dir_ircnn_train = "samples/train_{}".format(tl.global_flag['mode'])
tl.files.exists_or_mkdir(save_dir_ircnn_train)
checkpoint_dir = "checkpoint" # checkpoint
tl.files.exists_or_mkdir(checkpoint_dir)
###====================== PRE-LOAD DATA ===========================###
train_hr_img_list = sorted(tl.files.load_file_list(path=config.TRAIN.hr_img_path, regx='.*.jpg', printable=False))
train_lr_img_list = sorted(tl.files.load_file_list(path=config.TRAIN.lr_img_path, regx='.*.jpg', printable=False))
valid_hr_img_list = sorted(tl.files.load_file_list(path=config.VALID.hr_img_path, regx='.*.jpg', printable=False))
valid_lr_img_list = sorted(tl.files.load_file_list(path=config.VALID.lr_img_path, regx='.*.jpg', printable=False))
final_hr_img_list_4 = sorted(tl.files.load_file_list(path=config.FINAL.hr_img_path_4, regx='.*.png', printable=False))
final_lr_img_list_4 = sorted(tl.files.load_file_list(path=config.FINAL.lr_img_path_4, regx='.*.png', printable=False))
final_hr_img_list_6 = sorted(tl.files.load_file_list(path=config.FINAL.hr_img_path_6, regx='.*.png', printable=False))
final_lr_img_list_6 = sorted(tl.files.load_file_list(path=config.FINAL.lr_img_path_6, regx='.*.png', printable=False))
train_hr_imgs = tl.vis.read_images(train_hr_img_list, path=config.TRAIN.hr_img_path, n_threads=16)
###========================== DEFINE MODEL ============================###
## train inference
t_image = tf.placeholder('float32', [None, None, None, 3], name='t_image')
t_target_image = tf.placeholder('float32', [None, None, None, 3], name='t_target_image')
net = IRCNN(t_image, is_train=True, reuse=False)
net.print_params(False)
net.print_layers()
## test inference
net_test = IRCNN(t_image, is_train=False, reuse=True)
# ###========================== DEFINE TRAIN OPS ==========================###
loss = tl.cost.mean_squared_error(net.outputs, t_target_image, is_mean=True)
net_vars = tl.layers.get_variables_with_name('IRCNN', True, True)
with tf.variable_scope('learning_rate'):
lr_v = tf.Variable(lr_init, trainable=False)
## IRCNN
optim = tf.train.AdamOptimizer(lr_v, beta1=beta1).minimize(loss, var_list=net_vars)
###========================== RESTORE MODEL =============================###
sess = tf.Session(config=tf.ConfigProto(allow_soft_placement=True, log_device_placement=False))
tl.layers.initialize_global_variables(sess)
if tl.files.load_and_assign_npz(sess=sess, name=checkpoint_dir + '/{}.npz'.format(tl.global_flag['mode']), network=net) is True:
print('Load Last Checkpoint to IRCNN Model.\n')
else:
print('There is no checkpoint, do not load the model.\n')
###============================= TRAINING ===============================###
final_hr_imgs = tl.vis.read_images(final_hr_img_list_6, path=config.FINAL.hr_img_path_6, n_threads=16)
sample_imgs_hr = tl.prepro.threading_data(final_hr_imgs, fn=normalize_img_noresize, is_random=False)
print('final HR sub-image:', sample_imgs_hr.shape, sample_imgs_hr.min(), sample_imgs_hr.max())
final_lr_imgs = tl.vis.read_images(final_lr_img_list_6, path=config.FINAL.lr_img_path_6, n_threads=16)
sample_imgs_lr = tl.prepro.threading_data(final_lr_imgs, fn=normalize_img_noresize, is_random=False)
print('final LR sub-image:', sample_imgs_lr.shape, sample_imgs_lr.min(), sample_imgs_lr.max())
tl.vis.save_images(sample_imgs_lr, [1, 1], save_dir_ircnn + '/_final_sample_lr.jpg')
tl.vis.save_images(sample_imgs_hr, [1, 1], save_dir_ircnn + '/_final_sample_hr.jpg')
### ========================= train IRCNN ========================= ###
err_final, out = sess.run([loss,net_test.outputs], {t_image: sample_imgs_lr, t_target_image: sample_imgs_hr})
logging.debug("[*] Epoch: [%2d/%2d], Square Error on noise 0.6 is %f" %(0, n_epoch, err_final))
print ("[*] save images")
tl.vis.save_images(out, [1, 1], save_dir_ircnn + '/train_%d.png' % 0)
for epoch in range(0, n_epoch + 1):
## update learning rate
if epoch != 0 and (epoch % decay_every == 0):
new_lr_decay = lr_decay**(epoch // decay_every)
sess.run(tf.assign(lr_v, lr_init * new_lr_decay))
log = " ** new learning rate: %f " % (lr_init * new_lr_decay)
print(log)
elif epoch == 0:
sess.run(tf.assign(lr_v, lr_init))
log = " ** init lr: %f decay_every_init: %d, lr_decay: %f (for GAN)" % (lr_init, decay_every, lr_decay)
print(log)
epoch_time = time.time()
total_loss, n_iter = 0, 0
## If your machine have enough memory, please pre-load the whole train set.
for idx in range(0, len(train_hr_imgs), batch_size):
step_time = time.time()
b_imgs_hr = tl.prepro.threading_data(train_hr_imgs[idx:idx + batch_size], fn=normalize_img, is_random=True)
b_imgs_lr = tl.prepro.threading_data(train_hr_imgs[idx:idx + batch_size], fn=normalize_img_add_noise, noiseRatio=0.6)
## update IRCNN
err, out, _ = sess.run([loss, net.outputs, optim], {t_image: b_imgs_lr, t_target_image: b_imgs_hr})
print("Epoch [%2d/%2d] %4d time: %4.4fs, loss: %.8f" %
(epoch, n_epoch, n_iter, time.time() - step_time, err))
total_loss += err
n_iter += 1
log = "[*] Epoch: [%2d/%2d] time: %4.4fs, loss: %.8f" % (epoch, n_epoch, time.time() - epoch_time,
total_loss / n_iter)
print(log)
## quick evaluation on final set
if (epoch != 0) and (epoch % 5 == 0):
err_final, out = sess.run([loss,net_test.outputs], {t_image: sample_imgs_lr, t_target_image: sample_imgs_hr})
logging.debug("[*] Epoch: [%2d/%2d], Square Error on noise 0.6 is %f" %(epoch, n_epoch, err_final))
print("[*] save images")
tl.vis.save_images(out, [1, 1], save_dir_ircnn + '/train_%d.png' % epoch)
## save model
if (epoch != 0) and (epoch % 5 == 0):
tl.files.save_npz(net.all_params, name=checkpoint_dir + '/{}.npz'.format(tl.global_flag['mode']), sess=sess)
def evaluate():
## create folders to save result images
save_dir_ircnn = "final_image"
tl.files.exists_or_mkdir(save_dir_ircnn)
checkpoint_dir = "checkpoint"
###====================== PRE-LOAD DATA ===========================###
valid_lr_img_list = ["demo/B.png"]
valid_lr_imgs = tl.vis.read_images(valid_lr_img_list, path='.', n_threads=32)
###========================== DEFINE MODEL ============================###
valid_lr_img = valid_lr_imgs[0]
valid_lr_img = (valid_lr_img / 127.5) - 1
rols, cols, channels = valid_lr_img.shape
valid_lr_img = np.reshape(valid_lr_img, (1, rols, cols, channels))
t_image = tf.placeholder('float32', [1, None, None, 3], name='input_image')
net = IRCNN(t_image, is_train=False, reuse=False)
###===================== RESTORE IRCNN AND TEST =========================###
sess = tf.Session(config=tf.ConfigProto(allow_soft_placement=True, log_device_placement=False))
tl.layers.initialize_global_variables(sess)
tl.files.load_and_assign_npz(sess=sess, name=checkpoint_dir + '/ircnn.npz', network=net)
out = sess.run(net.outputs, {t_image: valid_lr_img})
print("[*] save images")
tl.vis.save_images(out, [1, 1], save_dir_ircnn + '/final.png')
if __name__ == '__main__':
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('--mode', type=str, default='ircnn', help='ircnn, evaluate')
args = parser.parse_args()
tl.global_flag['mode'] = args.mode
if tl.global_flag['mode'] == 'ircnn':
train()
elif tl.global_flag['mode'] == 'eval':
evaluate()
else:
raise Exception("Unknown --mode")