-
Notifications
You must be signed in to change notification settings - Fork 0
/
modelt.py
114 lines (108 loc) · 5.04 KB
/
modelt.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
import tensorflow as tf
import keras
import keras.backend as K
from keras import layers
from keras.layers import Input
from keras.layers import Activation
from keras.layers import Conv2D
from keras.layers import BatchNormalization
from keras.layers import DepthwiseConv2D
from keras.layers import SeparableConv2D
from keras.layers import Lambda
from keras.layers import concatenate
from keras.layers import AveragePooling2D
from keras.layers import UpSampling2D
from keras.layers import Dropout
from keras.layers import add
from keras.layers import Softmax
from keras.optimizers import SGD,Adam
from keras.models import Model
def conv_block(inputs,conv_type,kernel,kernel_size,strides,padding='same',relu=True):
## custom conv layer
## may use separableconv
if conv_type == 'ds':
x = SeparableConv2D(kernel,kernel_size,padding=padding,strides=strides)(inputs)
else:
x = Conv2D(kernel,kernel_size,padding=padding,strides=strides)(inputs)
x = BatchNormalization()(x)
if relu:
x = Activation('relu')(x)
return x
def _res_bottleneck(inputs,filters,kernel,t,s,r=False):
## residual block
tchannel = K.int_shape(inputs)[-1]*t
x = conv_block(inputs,'conv',tchannel,(1,1),strides=(1,1))
x = DepthwiseConv2D(kernel,strides=(s,s),depth_multiplier=1,padding='same')(x)
x = BatchNormalization()(x)
x = Activation('relu')(x)
x = conv_block(x,'conv',filters,(1,1),strides=(1,1),padding='same',relu=False)
if r:
x = add([x,inputs])
return x
def bottleneck_block(inputs,filters,kernel,t,strides,n):
## custom residual block.. real one...
x = _res_bottleneck(inputs,filters,kernel,t,strides)
for i in range(1,n):
x = _res_bottleneck(x,filters,kernel,t,1,True)
return x
def pyramid_pooling_block(input_tensor,bin_sizes):
## use input_shape instead fixed shape.
_,h,w,c = input_tensor.shape
concat_list = [input_tensor]
#w = 64
#h = 32
for bin_size in bin_sizes:
#x = AveragePooling2D(pool_size=(w//bin_size,h//bin_size),strides=(w//bin_size,h//bin_size))(input_tensor)
#x = Conv2D(128,(3,3),strides=(2,2),padding='same')(x)
x = AveragePooling2D(pool_size=(h//bin_size,w//bin_size),strides=(h//bin_size,w//bin_size))(input_tensor)
x = Conv2D(int(c) // len(bin_sizes),(3,3),strides=(2,2),padding='same')(x)
#x = Lambda(lambda x: tf.image.resize_images(x,(w,h)))(x)
x = Lambda(lambda x: tf.image.resize_images(x,(h,w)))(x)
concat_list.append(x)
return concatenate(concat_list)
def bulid_model(input_shape=(2048, 1024, 3)):
input_layer = Input(shape=input_shape,name='input_layer')
lds_layer = conv_block(input_layer,'conv',32,(3,3),strides=(2,2))
lds_layer = conv_block(lds_layer,'ds',48,(3,3),strides=(2,2))
lds_layer = conv_block(lds_layer,'ds',64,(3,3),strides=(2,2))
gfe_layer = bottleneck_block(lds_layer,64,(3,3),t=6,strides=2,n=3)
gfe_layer = bottleneck_block(gfe_layer,96,(3,3),t=6,strides=2,n=3)
gfe_layer = bottleneck_block(gfe_layer,128,(3,3),t=6,strides=1,n=3)
#gfe_layer = pyramid_pooling_block(gfe_layer,[2,4,6,8])
gfe_layer = pyramid_pooling_block(gfe_layer,[1,2,3,6])
################################################################
## feature fushion #########
ff_layer1 = conv_block(lds_layer,'conv',128,(1,1),padding='same',strides=(1,1),relu=False)
ff_layer2 = UpSampling2D((4,4))(gfe_layer)
ff_layer2 = DepthwiseConv2D((3,3),strides=(1,1),depth_multiplier=1,padding='same')(ff_layer2)
ff_layer2 = BatchNormalization()(ff_layer2)
ff_layer2 = Activation('relu')(ff_layer2)
ff_layer2 = Conv2D(128,(1,1),strides=(1,1),padding='same',activation=None)(ff_layer2)
ff_final = add([ff_layer1,ff_layer2])
ff_final = BatchNormalization()(ff_final)
ff_final = Activation('relu')(ff_final)
###############################################################
### Classifier ###
classifier = SeparableConv2D(128,(3,3),padding='same',strides=(1,1),name='DSConv1_classifier')(ff_final)
classifier = BatchNormalization()(classifier)
classifier = Activation('relu')(classifier)
classifier = SeparableConv2D(128,(3,3),padding='same',strides=(1,1),name='DSConv2_classifier')(classifier)
classifier = BatchNormalization()(classifier)
classifier = Activation('relu')(classifier)
## num classes.....
classifier = conv_block(classifier,'conv',21,(1,1),strides=(1,1),padding='same',relu=False)
classifier = Dropout(0.3)(classifier)
classifier = UpSampling2D((8,8))(classifier)
#classifier = Activation('softmax')(classifier)
classifier = Softmax()(classifier)
############################################################
fast_scnn = Model(inputs=input_layer,outputs=classifier,name='Fast_SCNN')
#optimizer = SGD(momentum=0.9,lr=0.045)
#optimizer = SGD(momentum=0.9,lr=1e-3)
#fast_scnn.compile(loss='categorical_crossentropy',optimizer=optimizer,metrics=['accuracy'])
return fast_scnn
if __name__ == '__main__':
model = bulid_model()
model.summary()
layerss = model.layers
print(layerss[-1].output)