-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathresnet_net.py
180 lines (143 loc) · 6.75 KB
/
resnet_net.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
#!/usr/bin/python3
# __*__ coding: utf-8 __*__
'''
@Author: SimonWang00
@Os：Windows 10 x64
@Contact: [email protected]
@Software: PY PyCharm
@File: settings.py
@Time: 2020/12/15 15:15
'''
# Copyright 2020 The SimonWang00. All Rights Reserved.
#
# @Desc:
# 1).building 2 layers of resnet block;
# 2).building 3 layers of resnet block;
# 3).define my resnet networks;
# ==============================================================================
# LINT.IfChange
"""define my resnet package"""
import tensorflow as tf
from typing import Tuple, List
from tensorflow.keras.layers import Conv2D, BatchNormalization, MaxPooling2D, Activation
def identity_block(input_data:tf.Tensor,
kernel_size: int or 3,
filters: Tuple[int, int, int],
stage: str or int,
block: str,
use_bias=False, train_bn=True) -> tf.Tensor:
'''
building 2 layers of resnet block
The alias's identity_block is the block that has no conv layer at shortcut
Parameters
----------
input_data :
kernel_size : default 3, the kernel size of middle conv layer at main path
filters : list of integers, the nb_filters of 3 conv layer at main path
stage : str or int, current stage label, used for generating layer names
block : laber name, 'a','b'..., current block label, used for generating layer names
use_bias : is use bias
train_bn : is use BatchNormalization
Returns
-------
'''
filter1, filter2, filter3 = filters
conv_name_base = 'res' + str(stage) + block + '_branch'
bn_name_base = 'bn' + str(stage) + block + '_branch'
x = Conv2D(filter1, (1, 1), padding="SAME", name=conv_name_base + '2a', use_bias=use_bias)(input_data)
x = BatchNormalization(name=bn_name_base + '2a')(x, training=train_bn)
x = Activation('relu')(x)
x = Conv2D(filter2, (kernel_size, kernel_size), padding="SAME", name=conv_name_base + '2b', use_bias=use_bias)(x)
x = BatchNormalization(name=bn_name_base + '2b')(x, training=train_bn)
x = Activation('relu')(x)
x = Conv2D(filter3, (1, 1), padding="SAME", name=conv_name_base + '2c', use_bias=use_bias)(x)
x = BatchNormalization(name=bn_name_base + '2c')(x, training=train_bn)
x = tf.keras.layers.Add()([x, input_data])
x = Activation('relu', name='res' + str(stage) + block + '_out')(x)
return x
def resnet_block_3(input_data,
kernel_size: int or 3,
filters: Tuple[int, int, int],
stage: str or int,
block: str,
strides=(2, 2) or Tuple[int, int],
use_bias=False,
train_bn=True) -> tf.Tensor:
'''
building 3 layers of resnet block
projection shortcut
Parameters
----------
input_data :
kernel_size :default 3, the kernel size of middle conv layer at main path
filters : list of integers, the nb_filters of 3 conv layer at main path
stage : current stage label, used for generating layer names
block : 'a','b'..., current block label, used for generating layer names
strides : the shortcut should have subsample=(2,2)
use_bias :
train_bn :
Returns
Note that from stage 3, the first conv layer at main path is with subsample=(2,2)
And the shortcut should have subsample=(2,2) as well
-------
'''
filter1, filter2, filter3 = filters
conv_name_base = 'res' + str(stage) + block + '_branch'
bn_name_base = 'bn' + str(stage) + block + '_branch'
x = Conv2D(filter1, (1, 1), strides=strides, padding="SAME", name=conv_name_base + '2a', use_bias=use_bias)(input_data)
x = BatchNormalization(name=bn_name_base + '2a')(x, training=train_bn)
x = Activation('relu')(x)
x = Conv2D(filter2, (kernel_size, kernel_size), padding="SAME", name=conv_name_base + '2b', use_bias=use_bias)(x)
x = BatchNormalization(name=bn_name_base + '2b')(x, training=train_bn)
x = Activation('relu')(x)
x = Conv2D(filter3, (1, 1), padding="SAME", name=conv_name_base +'2c', use_bias=use_bias)(x)
x = BatchNormalization(name=bn_name_base + '2c')(x, training=train_bn)
shortcut = Conv2D(filter3, (1, 1), padding="SAME", strides=strides, name=conv_name_base + '1', use_bias=use_bias)(input_data)
shortcut = BatchNormalization(name=bn_name_base + '1')(shortcut, training=train_bn)
x = tf.keras.layers.Add()([x, shortcut])
x = Activation('relu', name='res' + str(stage) + block + '_out')(x)
return x
def build_resnet(input_image,
architecture: str,
stage5=False, train_bn=True ) -> List:
'''
Build the CRNN network
Parameters
----------
input_image : inputs
architecture : Can be resnet50 or resnet101
stage5 : Boolean, If False, stage5 of the network is not created
train_bn : Boolean, is use BatchNormalization
Returns
-------
'''
assert architecture in ["resnet50", "resnet101"]
# Stage 1: input(b, 300, 1000, 3) -> output(b, 75, 250, 64)
x = tf.keras.layers.ZeroPadding2D((3, 3))(input_image)
x = Conv2D(64, (7, 7), strides=(2, 2), name='conv1', use_bias=True)(x)
x = BatchNormalization(name='bn_conv1')(x, training=train_bn)
x = Activation('relu')(x)
C1 = x = MaxPooling2D((3, 3), strides=(2, 2), padding="SAME")(x)
# Stage 2: input(b, 75, 250, 64) -> output(b, 75, 250, 256)
x = resnet_block_3(x, 3, (64, 64, 256), stage=2, block='a', strides=(1, 1), train_bn=train_bn)
x = identity_block(x, 3, (64, 64, 256), stage=2, block='b', train_bn=train_bn)
C2 = x = identity_block(x, 3, (64, 64, 256), stage=2, block='c', train_bn=train_bn)
# Stage 3: input(b, 75, 250, 256) -> output(b, 38, 125, 512)
x = resnet_block_3(x, 3, (128, 128, 512), stage=3, block='a', train_bn=train_bn)
x = identity_block(x, 3, (128, 128, 512), stage=3, block='b', train_bn=train_bn)
x = identity_block(x, 3, (128, 128, 512), stage=3, block='c', train_bn=train_bn)
C3 = x = identity_block(x, 3, (128, 128, 512), stage=3, block='d', train_bn=train_bn)
# Stage 4: input(b, 38, 125, 512) -> output(b, 19, 63, 1024)
x = resnet_block_3(x, 3, (256, 256, 1024), stage=4, block='a', train_bn=train_bn)
block_count = {"resnet50": 5, "resnet101": 22}[architecture]
for i in range(block_count):
x = identity_block(x, 3, (256, 256, 1024), stage=4, block=chr(98 + i), train_bn=train_bn)
C4 = x
# Stage 5: input(b, 19, 63, 1024) -> output(b, 19, 63, 2048)
if stage5:
x = resnet_block_3(x, 3, (512, 512, 2048), stage=5, block='a', train_bn=train_bn)
x = identity_block(x, 3, (512, 512, 2048), stage=5, block='b', train_bn=train_bn)
C5 = x = identity_block(x, 3, (512, 512, 2048), stage=5, block='c', train_bn=train_bn)
else:
C5 = None
return [C1, C2, C3, C4, C5]