-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsqueezenet.py
157 lines (141 loc) · 6.97 KB
/
squeezenet.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
from tensorflow.python.keras.models import Model
from tensorflow.python.keras.layers import Input, Activation, Concatenate
from tensorflow.python.keras.layers import Flatten, Dropout
from tensorflow.python.keras.layers import Convolution2D, MaxPooling2D
from tensorflow.python.keras.layers import GlobalAveragePooling2D
def SqueezeNet(nb_classes=8, inputs=(192, 192, 3), data_format='channels_last'):
""" Keras Implementation of SqueezeNet(arXiv 1602.07360)
@param nb_classes: total number of final categories
Arguments:
inputs -- shape of the input images (channel, cols, rows)
Source code taken from the following repository:
https://github.com/chasingbob/squeezenet-keras
"""
input_img = Input(shape=inputs)
conv1 = Convolution2D(
96, (7, 7), activation='relu', kernel_initializer='he_normal',
strides=(2, 2), padding='same', name='conv1',
data_format=data_format)(input_img)
maxpool1 = MaxPooling2D(
pool_size=(3, 3), strides=(2, 2), name='maxpool1',
data_format=data_format)(conv1)
fire2_squeeze = Convolution2D(
16, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire2_squeeze',
data_format=data_format)(maxpool1)
fire2_expand1 = Convolution2D(
64, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire2_expand1',
data_format=data_format)(fire2_squeeze)
fire2_expand2 = Convolution2D(
64, (3, 3), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire2_expand2',
data_format=data_format)(fire2_squeeze)
merge2 = Concatenate(axis=1)([fire2_expand1, fire2_expand2])
fire3_squeeze = Convolution2D(
16, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire3_squeeze',
data_format=data_format)(merge2)
fire3_expand1 = Convolution2D(
64, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire3_expand1',
data_format=data_format)(fire3_squeeze)
fire3_expand2 = Convolution2D(
64, (3, 3), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire3_expand2',
data_format=data_format)(fire3_squeeze)
merge3 = Concatenate(axis=1)([fire3_expand1, fire3_expand2])
fire4_squeeze = Convolution2D(
32, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire4_squeeze',
data_format=data_format)(merge3)
fire4_expand1 = Convolution2D(
128, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire4_expand1',
data_format=data_format)(fire4_squeeze)
fire4_expand2 = Convolution2D(
128, (3, 3), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire4_expand2',
data_format=data_format)(fire4_squeeze)
merge4 = Concatenate(axis=1)([fire4_expand1, fire4_expand2])
maxpool4 = MaxPooling2D(
pool_size=(3, 3), strides=(2, 2), name='maxpool4',
data_format=data_format)(merge4)
fire5_squeeze = Convolution2D(
32, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire5_squeeze',
data_format=data_format)(maxpool4)
fire5_expand1 = Convolution2D(
128, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire5_expand1',
data_format=data_format)(fire5_squeeze)
fire5_expand2 = Convolution2D(
128, (3, 3), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire5_expand2',
data_format=data_format)(fire5_squeeze)
merge5 = Concatenate(axis=1)([fire5_expand1, fire5_expand2])
fire6_squeeze = Convolution2D(
48, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire6_squeeze',
data_format=data_format)(merge5)
fire6_expand1 = Convolution2D(
192, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire6_expand1',
data_format=data_format)(fire6_squeeze)
fire6_expand2 = Convolution2D(
192, (3, 3), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire6_expand2',
data_format=data_format)(fire6_squeeze)
merge6 = Concatenate(axis=1)([fire6_expand1, fire6_expand2])
fire7_squeeze = Convolution2D(
48, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire7_squeeze',
data_format=data_format)(merge6)
fire7_expand1 = Convolution2D(
192, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire7_expand1',
data_format=data_format)(fire7_squeeze)
fire7_expand2 = Convolution2D(
192, (3, 3), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire7_expand2',
data_format=data_format)(fire7_squeeze)
merge7 = Concatenate(axis=1)([fire7_expand1, fire7_expand2])
fire8_squeeze = Convolution2D(
64, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire8_squeeze',
data_format=data_format)(merge7)
fire8_expand1 = Convolution2D(
256, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire8_expand1',
data_format=data_format)(fire8_squeeze)
fire8_expand2 = Convolution2D(
256, (3, 3), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire8_expand2',
data_format=data_format)(fire8_squeeze)
merge8 = Concatenate(axis=1)([fire8_expand1, fire8_expand2])
maxpool8 = MaxPooling2D(
pool_size=(3, 3), strides=(2, 2), name='maxpool8',
data_format=data_format)(merge8)
fire9_squeeze = Convolution2D(
64, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire9_squeeze',
data_format=data_format)(maxpool8)
fire9_expand1 = Convolution2D(
256, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire9_expand1',
data_format=data_format)(fire9_squeeze)
fire9_expand2 = Convolution2D(
256, (3, 3), activation='relu', kernel_initializer='he_normal',
padding='same', name='fire9_expand2',
data_format=data_format)(fire9_squeeze)
merge9 = Concatenate(axis=1)([fire9_expand1, fire9_expand2])
fire9_dropout = Dropout(0.5, name='fire9_dropout')(merge9)
conv10 = Convolution2D(
nb_classes, (1, 1), activation='relu', kernel_initializer='he_normal',
padding='valid', name='conv10',
data_format=data_format)(fire9_dropout)
global_avgpool10 = GlobalAveragePooling2D(data_format=data_format)(conv10)
softmax = Activation("softmax", name='softmax')(global_avgpool10)
return Model(inputs=input_img, outputs=softmax)
if __name__ == '__main__':
SqueezeNet().summary()