-
Notifications
You must be signed in to change notification settings - Fork 0
/
colorize.py
102 lines (78 loc) · 4.91 KB
/
colorize.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
import tensorflow as tf
import numpy as np
images = tf.placeholder(tf.float32, [None, 256, 256, 1])
# Conv 1
filters1_1 = tf.Variable(tf.truncated_normal([3, 3, 1, 64]))
bias1_1 = tf.Variable(tf.constant(0.1, shape=[256,256,64]))
conv1_1 = tf.relu(tf.nn.conv2d(images, filters1_1, strides=[1,1,1,1], padding='SAME') + bias1_1)
filters1_2 = tf.Variable(tf.truncated_normal([3, 3, 1, 64]))
bias1_2 = tf.Variable(tf.constant(0.1, shape=[128,128,64]))
conv1_2 = tf.relu(tf.nn.conv2d(conv1_1, filters1_2, strides=[2,2,1,1], padding='SAME') + bias1_2)
# Conv 2
filters2_1 = tf.Variable(tf.truncated_normal([3, 3, 1, 128]))
bias2_1 = tf.Variable(tf.constant(0.1, shape=[128,128,128]))
conv2_1 = tf.relu(tf.nn.conv2d(conv1_2, filters2_1, strides=[1,1,1,1], padding='SAME') + bias2_1)
filters2_2 = tf.Variable(tf.truncated_normal([3, 3, 1, 128]))
bias2_2 = tf.Variable(tf.constant(0.1, shape=[64,64,128]))
conv2_2 = tf.relu(tf.nn.conv2d(conv2_1, filters2_2, strides=[2,2,1,1], padding='SAME') + bias2_2)
# Conv 3
filters3_1 = tf.Variable(tf.truncated_normal([3, 3, 1, 256]))
bias3_1 = tf.Variable(tf.constant(0.1, shape=[64,64,256]))
conv3_1 = tf.relu(tf.nn.conv2d(conv2_2, filters3_1, strides=[1,1,1,1], padding='SAME') + bias3_1)
filters3_2 = tf.Variable(tf.truncated_normal([3, 3, 1, 256]))
bias3_2 = tf.Variable(tf.constant(0.1, shape=[64,64,256]))
conv3_2 = tf.relu(tf.nn.conv2d(conv3_1, filters3_2, strides=[1,1,1,1], padding='SAME') + bias3_2)
filters3_3 = tf.Variable(tf.truncated_normal([3, 3, 1, 256]))
bias3_3 = tf.Variable(tf.constant(0.1, shape=[32,32,256]))
conv3_3 = tf.relu(tf.nn.conv2d(conv3_2, filters3_3, strides=[2,2,1,1], padding='SAME') + bias3_3)
# Conv 4
filters4_1 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias4_1 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv4_1 = tf.relu(tf.nn.conv2d(conv3_3, filters4_1, strides=[1,1,1,1], padding='SAME') + bias4_1)
filters4_2 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias4_2 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv4_2 = tf.relu(tf.nn.conv2d(conv4_1, filters4_2, strides=[1,1,1,1], padding='SAME') + bias4_2)
filters4_3 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias4_3 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv4_3 = tf.relu(tf.nn.conv2d(conv4_2, filters4_3, strides=[1,1,1,1], padding='SAME') + bias4_3)
# Conv 5
filters5_1 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias5_1 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv5_1 = tf.relu(tf.nn.conv2d(conv4_3, filters5_1, strides=[1,1,1,1], padding='SAME') + bias5_1)
filters5_2 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias5_2 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv5_2 = tf.relu(tf.nn.conv2d(conv5_1, filters5_2, strides=[1,1,1,1], padding='SAME') + bias5_2)
filters5_3 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias5_3 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv5_3 = tf.relu(tf.nn.conv2d(conv5_2, filters5_3, strides=[1,1,1,1], padding='SAME') + bias5_3)
# Conv 6
filters6_1 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias6_1 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv6_1 = tf.relu(tf.nn.conv2d(conv5_3, filters6_1, strides=[1,1,1,1], padding='SAME') + bias6_1)
filters6_2 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias6_2 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv6_2 = tf.relu(tf.nn.conv2d(conv6_1, filters6_2, strides=[1,1,1,1], padding='SAME') + bias6_2)
filters6_3 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias6_3 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv6_3 = tf.relu(tf.nn.conv2d(conv6_2, filters6_3, strides=[1,1,1,1], padding='SAME') + bias6_3)
# Conv 7
filters7_1 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias7_1 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv7_1 = tf.relu(tf.nn.conv2d(conv6_3, filters7_1, strides=[1,1,1,1], padding='SAME') + bias7_1)
filters7_2 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias7_2 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv7_2 = tf.relu(tf.nn.conv2d(conv7_1, filters7_2, strides=[1,1,1,1], padding='SAME') + bias7_2)
filters7_3 = tf.Variable(tf.truncated_normal([3, 3, 1, 512]))
bias7_3 = tf.Variable(tf.constant(0.1, shape=[32,32,512]))
conv7_3 = tf.relu(tf.nn.conv2d(conv7_2, filters7_3, strides=[1,1,1,1], padding='SAME') + bias7_3)
# Conv 8
filters8_1 = tf.Variable(tf.truncated_normal([4, 4, 1, 256]))
bias8_1 = tf.Variable(tf.constant(0.1, shape=[64,64,256]))
conv8_1 = tf.relu(tf.nn.conv2d(conv7_3, filters8_1, strides=[2,2,1,1], padding='SAME') + bias8_1)
filters8_2 = tf.Variable(tf.truncated_normal([3, 3, 1, 256]))
bias8_2 = tf.Variable(tf.constant(0.1, shape=[64,64,256]))
conv8_2 = tf.relu(tf.nn.conv2d(conv8_1, filters8_2, strides=[1,1,1,1], padding='SAME') + bias8_2)
filters8_3 = tf.Variable(tf.truncated_normal([3, 3, 1, 256]))
bias8_3 = tf.Variable(tf.constant(0.1, shape=[64,64,256]))
conv8_3 = tf.relu(tf.nn.conv2d(conv8_2, filters8_3, strides=[1,1,1,1], padding='SAME') + bias8_3)
# Soft Max