CNN 常用模型
- 1. VGG16
-
- 2 Mobile_V1
- 3 Mobile_V2
- 4 Mobile_V3
1. VGG16
1.1 要点
- 13个Con2d(kernel_size=3) + 5个MaxPooling(pool_size=2,strides=2)+3个Fc
- 第一Block有2个kernel=3x3，filters=64的Conv2D,1个MaxPool。
- 第二Block有2个kernel=3x3，filters=128的Conv2D,1个MaxPool。
- 第三Block有3个kernel=3x3，filters=256的Conv2D,1个MaxPool。
- 第四Block有3个kernel=3x3，filters=512的Conv2D,1个MaxPool。
- 第五Block有3个kernel=3x3，filters=512的Conv2D,1个MaxPool。
- 第六Block有3个Fc，1个SoftMax。
- 网络结构如下
inputs(224x224x3)
↓
Conv-63(f=63,k=3,s=1,p='s',a='relu')
Conv-63
MaxPool
↓
Conv-128(f=128,k=3,s=1,p='s',a='relu')
Conv-128
MaxPool
↓
Conv-256(f=256,k=3,s=1,p='s',a='relu')
Conv-256
Conv-256
MaxPool
↓
Conv-512(f=512,k=3,s=1,p='s',a='relu')
Conv-512
Conv-512
MaxPool
↓
Conv-512(f=512,k=3,s=1,p='s',a='relu')
Conv-512
Conv-512
MaxPool
↓
Fc(4096)(u=4096,a='relu')
Fc(4096)
Fc(1000)(u=1000)
SoftMax
1.2 流程
- 导入程序包
- 编写Converlution blocks
- 编写Dense layers
- 建立model
1.3 代码
from keras import Model
from keras.utils import plot_model
from keras.layers import Input,Conv2D,BatchNormalization,MaxPool2D,Flatten,Dense,Dropout
input = Input(shape=(224,224,3))
x = Conv2D(filters=64,kernel_size=3,padding='same',activation='relu')(input)
x = Conv2D(filters=64,kernel_size=3,padding='same',activation='relu')(x)
x = MaxPool2D(pool_size=2,strides=2,padding='same')(x)
x = Conv2D(filters=128,kernel_size=3,padding='same',activation='relu')(x)
x = Conv2D(filters=128,kernel_size=3,padding='same',activation='relu')(x)
x = MaxPool2D(pool_size=2,strides=2,padding='same')(x)
x = Conv2D(filters=256,kernel_size=3,padding='same',activation='relu')(x)
x = Conv2D(filters=256,kernel_size=3,padding='same',activation='relu')(x)
x = Conv2D(filters=256,kernel_size=3,padding='same',activation='relu')(x)
x = MaxPool2D(pool_size=2,strides=2,padding='same')(x)
x = Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')(x)
x = Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')(x)
x = Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')(x)
x = MaxPool2D(pool_size=2,strides=2,padding='same')(x)
x = Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')(x)
x = Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')(x)
x = Conv2D(filters=512,kernel_size=3,padding='same',activation='relu')(x)
x = MaxPool2D(pool_size=2,strides=2,padding='same')(x)
x = Flatten()(x)
x = Dense(units = 4096,activation='relu')(x)
x = Dense(units = 4096,activation='relu')(x)
output = Dense(units = 1000,activation='softmax')(x)
model = Model(inputs=input,outputs=output)
model.summary()
plot_model(model,show_shapes = True)
_________________________________________________________________
Layer (type) Output Shape Param
=================================================================
input_1 (InputLayer) (None, 224, 224, 3) 0
_________________________________________________________________
conv2d_1 (Conv2D) (None, 224, 224, 64) 1792
_________________________________________________________________
conv2d_2 (Conv2D) (None, 224, 224, 64) 36928
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 112, 112, 64) 0
_________________________________________________________________
conv2d_3 (Conv2D) (None, 112, 112, 128) 73856
_________________________________________________________________
conv2d_4 (Conv2D) (None, 112, 112, 128) 147584
_________________________________________________________________
max_pooling2d_2 (MaxPooling2 (None, 56, 56, 128) 0
_________________________________________________________________
conv2d_5 (Conv2D) (None, 56, 56, 256) 295168
_________________________________________________________________
conv2d_6 (Conv2D) (None, 56, 56, 256) 590080
_________________________________________________________________
conv2d_7 (Conv2D) (None, 56, 56, 256) 590080
_________________________________________________________________
max_pooling2d_3 (MaxPooling2 (None, 28, 28, 256) 0
_________________________________________________________________
conv2d_8 (Conv2D) (None, 28, 28, 512) 1180160
_________________________________________________________________
conv2d_9 (Conv2D) (None, 28, 28, 512) 2359808
_________________________________________________________________
conv2d_10 (Conv2D) (None, 28, 28, 512) 2359808
_________________________________________________________________
max_pooling2d_4 (MaxPooling2 (None, 14, 14, 512) 0
_________________________________________________________________
conv2d_11 (Conv2D) (None, 14, 14, 512) 2359808
_________________________________________________________________
conv2d_12 (Conv2D) (None, 14, 14, 512) 2359808
_________________________________________________________________
conv2d_13 (Conv2D) (None, 14, 14, 512) 2359808
_________________________________________________________________
max_pooling2d_5 (MaxPooling2 (None, 7, 7, 512) 0
_________________________________________________________________
flatten_1 (Flatten) (None, 25088) 0
_________________________________________________________________
dense_1 (Dense) (None, 4096) 102764544
_________________________________________________________________
dense_2 (Dense) (None, 4096) 16781312
_________________________________________________________________
dense_3 (Dense) (None, 1000) 4097000
=================================================================
Total params: 138,357,544
Trainable params: 138,357,544
Non-trainable params: 0
_________________________________________________________________
Process finished with exit code 0
2 Mobile_V1
from keras import backend as K
from keras.models import Model
from keras.layers import DepthwiseConv2D,Input,Activation,Dropout,Reshape,BatchNormalization,GlobalAveragePooling2D,Conv2D
def _conv_block(inputs, filters, kernel=(3, 3), strides=(1, 1)):
'''conv2D +BatchNormalization + Activation '''
x = Conv2D(filters, kernel,
padding='same',
use_bias=False,
strides=strides,
name='conv1')(inputs)
x = BatchNormalization(name='conv1_bn')(x)
return Activation(relu6, name='conv1_relu')(x)
def _depthwise_conv_block(inputs, pointwise_conv_filters,
depth_multiplier=1, strides=(1, 1), block_id=1):
'''DepthwiseConv2D+BatchNormalization + Activation +conv2D + B + A'''
x = DepthwiseConv2D((3, 3),
padding='same',
depth_multiplier=depth_multiplier,
strides=strides,
use_bias=False,
name='conv_dw_%d' % block_id)(inputs)
x = BatchNormalization(name='conv_dw_%d_bn' % block_id)(x)
x = Activation(relu6, name='conv_dw_%d_relu' % block_id)(x)
x = Conv2D(pointwise_conv_filters, (1, 1),
padding='same',
use_bias=False,
strides=(1, 1),
name='conv_pw_%d' % block_id)(x)
x = BatchNormalization(name='conv_pw_%d_bn' % block_id)(x)
return Activation(relu6, name='conv_pw_%d_relu' % block_id)(x)
def relu6(x):
return K.relu(x, max_value=6)
def preprocess_input(x):
x /= 255.
x -= 0.5
x *= 2.
return x
def MobileNet(input_shape=[224,224,3],
depth_multiplier=1,
dropout=1e-3,
classes=1000):
img_input = Input(shape=input_shape)
x = _conv_block(img_input, 32, strides=(2, 2))
x = _depthwise_conv_block(x, 64, depth_multiplier, block_id=1)
x = _depthwise_conv_block(x, 128, depth_multiplier,
strides=(2, 2), block_id=2)
x = _depthwise_conv_block(x, 128, depth_multiplier, block_id=3)
x = _depthwise_conv_block(x, 256, depth_multiplier,
strides=(2, 2), block_id=4)
x = _depthwise_conv_block(x, 256, depth_multiplier, block_id=5)
x = _depthwise_conv_block(x, 512, depth_multiplier,
strides=(2, 2), block_id=6)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=7)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=8)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=9)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=10)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=11)
x = _depthwise_conv_block(x, 1024, depth_multiplier,
strides=(2, 2), block_id=12)
x = _depthwise_conv_block(x, 1024, depth_multiplier, block_id=13)
x = GlobalAveragePooling2D()(x)
x = Reshape((1, 1, 1024), name='reshape_1')(x)
x = Dropout(dropout, name='dropout')(x)
x = Conv2D(classes, (1, 1),padding='same', name='conv_preds')(x)
x = Activation('softmax', name='act_softmax')(x)
x = Reshape((classes,), name='reshape_2')(x)
inputs = img_input
model = Model(inputs, x)
return model
if __name__ == '__main__':
model = MobileNet(input_shape=(224, 224, 3))
model.summary()
'''
inputs:[224,224,3 ]
outputs:[1,1,000]
[224,224,3 ]-->_conv_block + _depthwise_conv_block-->
[112,112,64]-->_depthwise_conv_block*2 -->
[56,56,128]-->_depthwise_conv_block*2 -->
[28,28,256]-->_depthwise_conv_block-->
[14,14,512]-->_depthwise_conv_block*5-->
[14,14,512]-->_depthwise_conv_block*2-->
[ 7,7,1024]-->GlobalAveragePooling2D + Dropout-->
[ 1,1,1024]-->Conv2D + Activation-->[ 1, 1, 1000]
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_1 (InputLayer) (None, 224, 224, 3) 0
_________________________________________________________________
conv1 (Conv2D) (None, 112, 112, 32) 864
_________________________________________________________________
conv1_bn (BatchNormalization (None, 112, 112, 32) 128
_________________________________________________________________
conv1_relu (Activation) (None, 112, 112, 32) 0
_________________________________________________________________
conv_dw_1 (DepthwiseConv2D) (None, 112, 112, 32) 288
_________________________________________________________________
conv_dw_1_bn (BatchNormaliza (None, 112, 112, 32) 128
_________________________________________________________________
conv_dw_1_relu (Activation) (None, 112, 112, 32) 0
_________________________________________________________________
conv_pw_1 (Conv2D) (None, 112, 112, 64) 2048
_________________________________________________________________
conv_pw_1_bn (BatchNormaliza (None, 112, 112, 64) 256
_________________________________________________________________
conv_pw_1_relu (Activation) (None, 112, 112, 64) 0
_________________________________________________________________
conv_dw_2 (DepthwiseConv2D) (None, 56, 56, 64) 576
_________________________________________________________________
conv_dw_2_bn (BatchNormaliza (None, 56, 56, 64) 256
_________________________________________________________________
conv_dw_2_relu (Activation) (None, 56, 56, 64) 0
_________________________________________________________________
conv_pw_2 (Conv2D) (None, 56, 56, 128) 8192
_________________________________________________________________
conv_pw_2_bn (BatchNormaliza (None, 56, 56, 128) 512
_________________________________________________________________
conv_pw_2_relu (Activation) (None, 56, 56, 128) 0
_________________________________________________________________
conv_dw_3 (DepthwiseConv2D) (None, 56, 56, 128) 1152
_________________________________________________________________
conv_dw_3_bn (BatchNormaliza (None, 56, 56, 128) 512
_________________________________________________________________
conv_dw_3_relu (Activation) (None, 56, 56, 128) 0
_________________________________________________________________
conv_pw_3 (Conv2D) (None, 56, 56, 128) 16384
_________________________________________________________________
conv_pw_3_bn (BatchNormaliza (None, 56, 56, 128) 512
_________________________________________________________________
conv_pw_3_relu (Activation) (None, 56, 56, 128) 0
_________________________________________________________________
conv_dw_4 (DepthwiseConv2D) (None, 28, 28, 128) 1152
_________________________________________________________________
conv_dw_4_bn (BatchNormaliza (None, 28, 28, 128) 512
_________________________________________________________________
conv_dw_4_relu (Activation) (None, 28, 28, 128) 0
_________________________________________________________________
conv_pw_4 (Conv2D) (None, 28, 28, 256) 32768
_________________________________________________________________
conv_pw_4_bn (BatchNormaliza (None, 28, 28, 256) 1024
_________________________________________________________________
conv_pw_4_relu (Activation) (None, 28, 28, 256) 0
_________________________________________________________________
conv_dw_5 (DepthwiseConv2D) (None, 28, 28, 256) 2304
_________________________________________________________________
conv_dw_5_bn (BatchNormaliza (None, 28, 28, 256) 1024
_________________________________________________________________
conv_dw_5_relu (Activation) (None, 28, 28, 256) 0
_________________________________________________________________
conv_pw_5 (Conv2D) (None, 28, 28, 256) 65536
_________________________________________________________________
conv_pw_5_bn (BatchNormaliza (None, 28, 28, 256) 1024
_________________________________________________________________
conv_pw_5_relu (Activation) (None, 28, 28, 256) 0
_________________________________________________________________
conv_dw_6 (DepthwiseConv2D) (None, 14, 14, 256) 2304
_________________________________________________________________
conv_dw_6_bn (BatchNormaliza (None, 14, 14, 256) 1024
_________________________________________________________________
conv_dw_6_relu (Activation) (None, 14, 14, 256) 0
_________________________________________________________________
conv_pw_6 (Conv2D) (None, 14, 14, 512) 131072
_________________________________________________________________
conv_pw_6_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_6_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_7 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_7_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_7_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_7 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_7_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_7_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_8 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_8_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_8_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_8 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_8_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_8_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_9 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_9_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_9_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_9 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_9_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_9_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_10 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_10_bn (BatchNormaliz (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_10_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_10 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_10_bn (BatchNormaliz (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_10_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_11 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_11_bn (BatchNormaliz (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_11_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_11 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_11_bn (BatchNormaliz (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_11_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_12 (DepthwiseConv2D) (None, 7, 7, 512) 4608
_________________________________________________________________
conv_dw_12_bn (BatchNormaliz (None, 7, 7, 512) 2048
_________________________________________________________________
conv_dw_12_relu (Activation) (None, 7, 7, 512) 0
_________________________________________________________________
conv_pw_12 (Conv2D) (None, 7, 7, 1024) 524288
_________________________________________________________________
conv_pw_12_bn (BatchNormaliz (None, 7, 7, 1024) 4096
_________________________________________________________________
conv_pw_12_relu (Activation) (None, 7, 7, 1024) 0
_________________________________________________________________
conv_dw_13 (DepthwiseConv2D) (None, 7, 7, 1024) 9216
_________________________________________________________________
conv_dw_13_bn (BatchNormaliz (None, 7, 7, 1024) 4096
_________________________________________________________________
conv_dw_13_relu (Activation) (None, 7, 7, 1024) 0
_________________________________________________________________
conv_pw_13 (Conv2D) (None, 7, 7, 1024) 1048576
_________________________________________________________________
conv_pw_13_bn (BatchNormaliz (None, 7, 7, 1024) 4096
_________________________________________________________________
conv_pw_13_relu (Activation) (None, 7, 7, 1024) 0
_________________________________________________________________
global_average_pooling2d_1 ( (None, 1024) 0
_________________________________________________________________
reshape_1 (Reshape) (None, 1, 1, 1024) 0
_________________________________________________________________
dropout (Dropout) (None, 1, 1, 1024) 0
_________________________________________________________________
conv_preds (Conv2D) (None, 1, 1, 1000) 1025000
_________________________________________________________________
act_softmax (Activation) (None, 1, 1, 1000) 0
_________________________________________________________________
reshape_2 (Reshape) (None, 1000) 0
=================================================================
Total params: 4,253,864
Trainable params: 4,231,976
Non-trainable params: 21,888
_________________________________________________________________
Process finished with exit code 0
'''
3 Mobile_V2
from keras import backend as K
from keras.models import Model
from keras.layers import DepthwiseConv2D,Input,Activation,Dropout,Reshape,BatchNormalization,GlobalAveragePooling2D,Conv2D
def _conv_block(inputs, filters, kernel=(3, 3), strides=(1, 1)):
'''conv2D +BatchNormalization + Activation '''
x = Conv2D(filters, kernel,
padding='same',
use_bias=False,
strides=strides,
name='conv1')(inputs)
x = BatchNormalization(name='conv1_bn')(x)
return Activation(relu6, name='conv1_relu')(x)
def _depthwise_conv_block(inputs, pointwise_conv_filters,
depth_multiplier=1, strides=(1, 1), block_id=1):
'''DepthwiseConv2D+BatchNormalization + Activation +conv2D + B + A'''
x = DepthwiseConv2D((3, 3),
padding='same',
depth_multiplier=depth_multiplier,
strides=strides,
use_bias=False,
name='conv_dw_%d' % block_id)(inputs)
x = BatchNormalization(name='conv_dw_%d_bn' % block_id)(x)
x = Activation(relu6, name='conv_dw_%d_relu' % block_id)(x)
x = Conv2D(pointwise_conv_filters, (1, 1),
padding='same',
use_bias=False,
strides=(1, 1),
name='conv_pw_%d' % block_id)(x)
x = BatchNormalization(name='conv_pw_%d_bn' % block_id)(x)
return Activation(relu6, name='conv_pw_%d_relu' % block_id)(x)
def relu6(x):
return K.relu(x, max_value=6)
def preprocess_input(x):
x /= 255.
x -= 0.5
x *= 2.
return x
def MobileNet(input_shape=[224,224,3],
depth_multiplier=1,
dropout=1e-3,
classes=1000):
img_input = Input(shape=input_shape)
x = _conv_block(img_input, 32, strides=(2, 2))
x = _depthwise_conv_block(x, 64, depth_multiplier, block_id=1)
x = _depthwise_conv_block(x, 128, depth_multiplier,
strides=(2, 2), block_id=2)
x = _depthwise_conv_block(x, 128, depth_multiplier, block_id=3)
x = _depthwise_conv_block(x, 256, depth_multiplier,
strides=(2, 2), block_id=4)
x = _depthwise_conv_block(x, 256, depth_multiplier, block_id=5)
x = _depthwise_conv_block(x, 512, depth_multiplier,
strides=(2, 2), block_id=6)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=7)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=8)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=9)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=10)
x = _depthwise_conv_block(x, 512, depth_multiplier, block_id=11)
x = _depthwise_conv_block(x, 1024, depth_multiplier,
strides=(2, 2), block_id=12)
x = _depthwise_conv_block(x, 1024, depth_multiplier, block_id=13)
x = GlobalAveragePooling2D()(x)
x = Reshape((1, 1, 1024), name='reshape_1')(x)
x = Dropout(dropout, name='dropout')(x)
x = Conv2D(classes, (1, 1),padding='same', name='conv_preds')(x)
x = Activation('softmax', name='act_softmax')(x)
x = Reshape((classes,), name='reshape_2')(x)
inputs = img_input
model = Model(inputs, x)
return model
if __name__ == '__main__':
model = MobileNet(input_shape=(224, 224, 3))
model.summary()
'''
inputs:[224,224,3 ]
outputs:[1,1,000]
[224,224,3 ]-->_conv_block + _depthwise_conv_block-->
[112,112,64]-->_depthwise_conv_block*2 -->
[56,56,128]-->_depthwise_conv_block*2 -->
[28,28,256]-->_depthwise_conv_block-->
[14,14,512]-->_depthwise_conv_block*5-->
[14,14,512]-->_depthwise_conv_block*2-->
[ 7,7,1024]-->GlobalAveragePooling2D + Dropout-->
[ 1,1,1024]-->Conv2D + Activation-->[ 1, 1, 1000]
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_1 (InputLayer) (None, 224, 224, 3) 0
_________________________________________________________________
conv1 (Conv2D) (None, 112, 112, 32) 864
_________________________________________________________________
conv1_bn (BatchNormalization (None, 112, 112, 32) 128
_________________________________________________________________
conv1_relu (Activation) (None, 112, 112, 32) 0
_________________________________________________________________
conv_dw_1 (DepthwiseConv2D) (None, 112, 112, 32) 288
_________________________________________________________________
conv_dw_1_bn (BatchNormaliza (None, 112, 112, 32) 128
_________________________________________________________________
conv_dw_1_relu (Activation) (None, 112, 112, 32) 0
_________________________________________________________________
conv_pw_1 (Conv2D) (None, 112, 112, 64) 2048
_________________________________________________________________
conv_pw_1_bn (BatchNormaliza (None, 112, 112, 64) 256
_________________________________________________________________
conv_pw_1_relu (Activation) (None, 112, 112, 64) 0
_________________________________________________________________
conv_dw_2 (DepthwiseConv2D) (None, 56, 56, 64) 576
_________________________________________________________________
conv_dw_2_bn (BatchNormaliza (None, 56, 56, 64) 256
_________________________________________________________________
conv_dw_2_relu (Activation) (None, 56, 56, 64) 0
_________________________________________________________________
conv_pw_2 (Conv2D) (None, 56, 56, 128) 8192
_________________________________________________________________
conv_pw_2_bn (BatchNormaliza (None, 56, 56, 128) 512
_________________________________________________________________
conv_pw_2_relu (Activation) (None, 56, 56, 128) 0
_________________________________________________________________
conv_dw_3 (DepthwiseConv2D) (None, 56, 56, 128) 1152
_________________________________________________________________
conv_dw_3_bn (BatchNormaliza (None, 56, 56, 128) 512
_________________________________________________________________
conv_dw_3_relu (Activation) (None, 56, 56, 128) 0
_________________________________________________________________
conv_pw_3 (Conv2D) (None, 56, 56, 128) 16384
_________________________________________________________________
conv_pw_3_bn (BatchNormaliza (None, 56, 56, 128) 512
_________________________________________________________________
conv_pw_3_relu (Activation) (None, 56, 56, 128) 0
_________________________________________________________________
conv_dw_4 (DepthwiseConv2D) (None, 28, 28, 128) 1152
_________________________________________________________________
conv_dw_4_bn (BatchNormaliza (None, 28, 28, 128) 512
_________________________________________________________________
conv_dw_4_relu (Activation) (None, 28, 28, 128) 0
_________________________________________________________________
conv_pw_4 (Conv2D) (None, 28, 28, 256) 32768
_________________________________________________________________
conv_pw_4_bn (BatchNormaliza (None, 28, 28, 256) 1024
_________________________________________________________________
conv_pw_4_relu (Activation) (None, 28, 28, 256) 0
_________________________________________________________________
conv_dw_5 (DepthwiseConv2D) (None, 28, 28, 256) 2304
_________________________________________________________________
conv_dw_5_bn (BatchNormaliza (None, 28, 28, 256) 1024
_________________________________________________________________
conv_dw_5_relu (Activation) (None, 28, 28, 256) 0
_________________________________________________________________
conv_pw_5 (Conv2D) (None, 28, 28, 256) 65536
_________________________________________________________________
conv_pw_5_bn (BatchNormaliza (None, 28, 28, 256) 1024
_________________________________________________________________
conv_pw_5_relu (Activation) (None, 28, 28, 256) 0
_________________________________________________________________
conv_dw_6 (DepthwiseConv2D) (None, 14, 14, 256) 2304
_________________________________________________________________
conv_dw_6_bn (BatchNormaliza (None, 14, 14, 256) 1024
_________________________________________________________________
conv_dw_6_relu (Activation) (None, 14, 14, 256) 0
_________________________________________________________________
conv_pw_6 (Conv2D) (None, 14, 14, 512) 131072
_________________________________________________________________
conv_pw_6_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_6_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_7 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_7_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_7_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_7 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_7_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_7_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_8 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_8_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_8_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_8 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_8_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_8_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_9 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_9_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_9_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_9 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_9_bn (BatchNormaliza (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_9_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_10 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_10_bn (BatchNormaliz (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_10_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_10 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_10_bn (BatchNormaliz (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_10_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_11 (DepthwiseConv2D) (None, 14, 14, 512) 4608
_________________________________________________________________
conv_dw_11_bn (BatchNormaliz (None, 14, 14, 512) 2048
_________________________________________________________________
conv_dw_11_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_pw_11 (Conv2D) (None, 14, 14, 512) 262144
_________________________________________________________________
conv_pw_11_bn (BatchNormaliz (None, 14, 14, 512) 2048
_________________________________________________________________
conv_pw_11_relu (Activation) (None, 14, 14, 512) 0
_________________________________________________________________
conv_dw_12 (DepthwiseConv2D) (None, 7, 7, 512) 4608
_________________________________________________________________
conv_dw_12_bn (BatchNormaliz (None, 7, 7, 512) 2048
_________________________________________________________________
conv_dw_12_relu (Activation) (None, 7, 7, 512) 0
_________________________________________________________________
conv_pw_12 (Conv2D) (None, 7, 7, 1024) 524288
_________________________________________________________________
conv_pw_12_bn (BatchNormaliz (None, 7, 7, 1024) 4096
_________________________________________________________________
conv_pw_12_relu (Activation) (None, 7, 7, 1024) 0
_________________________________________________________________
conv_dw_13 (DepthwiseConv2D) (None, 7, 7, 1024) 9216
_________________________________________________________________
conv_dw_13_bn (BatchNormaliz (None, 7, 7, 1024) 4096
_________________________________________________________________
conv_dw_13_relu (Activation) (None, 7, 7, 1024) 0
_________________________________________________________________
conv_pw_13 (Conv2D) (None, 7, 7, 1024) 1048576
_________________________________________________________________
conv_pw_13_bn (BatchNormaliz (None, 7, 7, 1024) 4096
_________________________________________________________________
conv_pw_13_relu (Activation) (None, 7, 7, 1024) 0
_________________________________________________________________
global_average_pooling2d_1 ( (None, 1024) 0
_________________________________________________________________
reshape_1 (Reshape) (None, 1, 1, 1024) 0
_________________________________________________________________
dropout (Dropout) (None, 1, 1, 1024) 0
_________________________________________________________________
conv_preds (Conv2D) (None, 1, 1, 1000) 1025000
_________________________________________________________________
act_softmax (Activation) (None, 1, 1, 1000) 0
_________________________________________________________________
reshape_2 (Reshape) (None, 1000) 0
=================================================================
Total params: 4,253,864
Trainable params: 4,231,976
Non-trainable params: 21,888
_________________________________________________________________
Process finished with exit code 0
'''
4 Mobile_V3
from keras.layers import Conv2D, DepthwiseConv2D, Dense, GlobalAveragePooling2D,Input
from keras.layers import Activation, BatchNormalization, Add, Multiply, Reshape
from keras.models import Model
from keras import backend as K
alpha = 1
def relu6(x):
return K.relu(x, max_value=6.0)
def hard_swish(x):
return x * K.relu(x + 3.0, max_value=6.0) / 6.0
def return_activation(x, nl):
if nl == 'HS':
x = Activation(hard_swish)(x)
if nl == 'RE':
x = Activation(relu6)(x)
return x
def conv_block(inputs, filters, kernel, strides, nl):
channel_axis = 1 if K.image_data_format() == 'channels_first' else -1
x = Conv2D(filters, kernel, padding='same', strides=strides)(inputs)
x = BatchNormalization(axis=channel_axis)(x)
return return_activation(x, nl)
def squeeze(inputs):
input_channels = int(inputs.shape[-1])
x = GlobalAveragePooling2D()(inputs)
x = Dense(int(input_channels/4))(x)
x = Activation(relu6)(x)
x = Dense(input_channels)(x)
x = Activation(hard_swish)(x)
x = Reshape((1, 1, input_channels))(x)
x = Multiply()([inputs, x])
return x
def bottleneck(inputs, filters, kernel, up_dim, stride, sq, nl):
channel_axis = 1 if K.image_data_format() == 'channels_first' else -1
input_shape = K.int_shape(inputs)
tchannel = int(up_dim)
cchannel = int(alpha * filters)
r = stride == 1 and input_shape[3] == filters
x = conv_block(inputs, tchannel, (1, 1), (1, 1), nl)
x = DepthwiseConv2D(kernel, strides=(stride, stride), depth_multiplier=1, padding='same')(x)
x = BatchNormalization(axis=channel_axis)(x)
x = return_activation(x, nl)
if sq:
x = squeeze(x)
x = Conv2D(cchannel, (1, 1), strides=(1, 1), padding='same')(x)
x = BatchNormalization(axis=channel_axis)(x)
if r:
x = Add()([x, inputs])
return x
def MobileNetv3_small(shape = (224,224,3),n_class = 1000):
inputs = Input(shape)
x = conv_block(inputs, 16, (3, 3), strides=(2, 2), nl='HS')
x = bottleneck(x, 16, (3, 3), up_dim=16, stride=2, sq=True, nl='RE')
x = bottleneck(x, 24, (3, 3), up_dim=72, stride=2, sq=False, nl='RE')
x = bottleneck(x, 24, (3, 3), up_dim=88, stride=1, sq=False, nl='RE')
x = bottleneck(x, 40, (5, 5), up_dim=96, stride=2, sq=True, nl='HS')
x = bottleneck(x, 40, (5, 5), up_dim=240, stride=1, sq=True, nl='HS')
x = bottleneck(x, 40, (5, 5), up_dim=240, stride=1, sq=True, nl='HS')
x = bottleneck(x, 48, (5, 5), up_dim=120, stride=1, sq=True, nl='HS')
x = bottleneck(x, 48, (5, 5), up_dim=144, stride=1, sq=True, nl='HS')
x = bottleneck(x, 96, (5, 5), up_dim=288, stride=2, sq=True, nl='HS')
x = bottleneck(x, 96, (5, 5), up_dim=576, stride=1, sq=True, nl='HS')
x = bottleneck(x, 96, (5, 5), up_dim=576, stride=1, sq=True, nl='HS')
x = conv_block(x, 576, (1, 1), strides=(1, 1), nl='HS')
x = GlobalAveragePooling2D()(x)
x = Reshape((1, 1, 576))(x)
x = Conv2D(1024, (1, 1), padding='same')(x)
x = return_activation(x, 'HS')
x = Conv2D(n_class, (1, 1), padding='same', activation='softmax')(x)
x = Reshape((n_class,))(x)
model = Model(inputs, x)
return model
if __name__ == "__main__":
model = MobileNetv3_small()
model.summary()
'''
inputs: [224, 224, 3]
outputs: [None, 1000]
[224, 224, 3]-->conv_block-->[112,112,16]
[112,112,16]--> bottleneck-->[56,56,1656,56,16]
[56,56,16]-->bottleneck*2 -->[28,28,24]
[28,28,24]-->bottleneck*3 -->[14,14,40]
[14,14,40]-->bottleneck*2 -->[14,14,48]
[14,14,48]-->bottleneck*3 -->[7,7,96]
[7,7,96]--> conv_block + GlobalAveragePooling2D-->[1, 1, 576]
[1, 1, 576]-->Conv2D -->[1, 1, 1024]
[1, 1, 1024]-->>Conv2D -->[None, 1000]
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
input_1 (InputLayer) (None, 224, 224, 3) 0
__________________________________________________________________________________________________
conv2d_1 (Conv2D) (None, 112, 112, 16) 448 input_1[0][0]
__________________________________________________________________________________________________
batch_normalization_1 (BatchNor (None, 112, 112, 16) 64 conv2d_1[0][0]
__________________________________________________________________________________________________
activation_1 (Activation) (None, 112, 112, 16) 0 batch_normalization_1[0][0]
__________________________________________________________________________________________________
conv2d_2 (Conv2D) (None, 112, 112, 16) 272 activation_1[0][0]
__________________________________________________________________________________________________
batch_normalization_2 (BatchNor (None, 112, 112, 16) 64 conv2d_2[0][0]
__________________________________________________________________________________________________
activation_2 (Activation) (None, 112, 112, 16) 0 batch_normalization_2[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_1 (DepthwiseCo (None, 56, 56, 16) 160 activation_2[0][0]
__________________________________________________________________________________________________
batch_normalization_3 (BatchNor (None, 56, 56, 16) 64 depthwise_conv2d_1[0][0]
__________________________________________________________________________________________________
activation_3 (Activation) (None, 56, 56, 16) 0 batch_normalization_3[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_1 (Glo (None, 16) 0 activation_3[0][0]
__________________________________________________________________________________________________
dense_1 (Dense) (None, 4) 68 global_average_pooling2d_1[0][0]
__________________________________________________________________________________________________
activation_4 (Activation) (None, 4) 0 dense_1[0][0]
__________________________________________________________________________________________________
dense_2 (Dense) (None, 16) 80 activation_4[0][0]
__________________________________________________________________________________________________
activation_5 (Activation) (None, 16) 0 dense_2[0][0]
__________________________________________________________________________________________________
reshape_1 (Reshape) (None, 1, 1, 16) 0 activation_5[0][0]
__________________________________________________________________________________________________
multiply_1 (Multiply) (None, 56, 56, 16) 0 activation_3[0][0]
reshape_1[0][0]
__________________________________________________________________________________________________
conv2d_3 (Conv2D) (None, 56, 56, 16) 272 multiply_1[0][0]
__________________________________________________________________________________________________
batch_normalization_4 (BatchNor (None, 56, 56, 16) 64 conv2d_3[0][0]
__________________________________________________________________________________________________
conv2d_4 (Conv2D) (None, 56, 56, 72) 1224 batch_normalization_4[0][0]
__________________________________________________________________________________________________
batch_normalization_5 (BatchNor (None, 56, 56, 72) 288 conv2d_4[0][0]
__________________________________________________________________________________________________
activation_6 (Activation) (None, 56, 56, 72) 0 batch_normalization_5[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_2 (DepthwiseCo (None, 28, 28, 72) 720 activation_6[0][0]
__________________________________________________________________________________________________
batch_normalization_6 (BatchNor (None, 28, 28, 72) 288 depthwise_conv2d_2[0][0]
__________________________________________________________________________________________________
activation_7 (Activation) (None, 28, 28, 72) 0 batch_normalization_6[0][0]
__________________________________________________________________________________________________
conv2d_5 (Conv2D) (None, 28, 28, 24) 1752 activation_7[0][0]
__________________________________________________________________________________________________
batch_normalization_7 (BatchNor (None, 28, 28, 24) 96 conv2d_5[0][0]
__________________________________________________________________________________________________
conv2d_6 (Conv2D) (None, 28, 28, 88) 2200 batch_normalization_7[0][0]
__________________________________________________________________________________________________
batch_normalization_8 (BatchNor (None, 28, 28, 88) 352 conv2d_6[0][0]
__________________________________________________________________________________________________
activation_8 (Activation) (None, 28, 28, 88) 0 batch_normalization_8[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_3 (DepthwiseCo (None, 28, 28, 88) 880 activation_8[0][0]
__________________________________________________________________________________________________
batch_normalization_9 (BatchNor (None, 28, 28, 88) 352 depthwise_conv2d_3[0][0]
__________________________________________________________________________________________________
activation_9 (Activation) (None, 28, 28, 88) 0 batch_normalization_9[0][0]
__________________________________________________________________________________________________
conv2d_7 (Conv2D) (None, 28, 28, 24) 2136 activation_9[0][0]
__________________________________________________________________________________________________
batch_normalization_10 (BatchNo (None, 28, 28, 24) 96 conv2d_7[0][0]
__________________________________________________________________________________________________
add_1 (Add) (None, 28, 28, 24) 0 batch_normalization_10[0][0]
batch_normalization_7[0][0]
__________________________________________________________________________________________________
conv2d_8 (Conv2D) (None, 28, 28, 96) 2400 add_1[0][0]
__________________________________________________________________________________________________
batch_normalization_11 (BatchNo (None, 28, 28, 96) 384 conv2d_8[0][0]
__________________________________________________________________________________________________
activation_10 (Activation) (None, 28, 28, 96) 0 batch_normalization_11[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_4 (DepthwiseCo (None, 14, 14, 96) 2496 activation_10[0][0]
__________________________________________________________________________________________________
batch_normalization_12 (BatchNo (None, 14, 14, 96) 384 depthwise_conv2d_4[0][0]
__________________________________________________________________________________________________
activation_11 (Activation) (None, 14, 14, 96) 0 batch_normalization_12[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_2 (Glo (None, 96) 0 activation_11[0][0]
__________________________________________________________________________________________________
dense_3 (Dense) (None, 24) 2328 global_average_pooling2d_2[0][0]
__________________________________________________________________________________________________
activation_12 (Activation) (None, 24) 0 dense_3[0][0]
__________________________________________________________________________________________________
dense_4 (Dense) (None, 96) 2400 activation_12[0][0]
__________________________________________________________________________________________________
activation_13 (Activation) (None, 96) 0 dense_4[0][0]
__________________________________________________________________________________________________
reshape_2 (Reshape) (None, 1, 1, 96) 0 activation_13[0][0]
__________________________________________________________________________________________________
multiply_2 (Multiply) (None, 14, 14, 96) 0 activation_11[0][0]
reshape_2[0][0]
__________________________________________________________________________________________________
conv2d_9 (Conv2D) (None, 14, 14, 40) 3880 multiply_2[0][0]
__________________________________________________________________________________________________
batch_normalization_13 (BatchNo (None, 14, 14, 40) 160 conv2d_9[0][0]
__________________________________________________________________________________________________
conv2d_10 (Conv2D) (None, 14, 14, 240) 9840 batch_normalization_13[0][0]
__________________________________________________________________________________________________
batch_normalization_14 (BatchNo (None, 14, 14, 240) 960 conv2d_10[0][0]
__________________________________________________________________________________________________
activation_14 (Activation) (None, 14, 14, 240) 0 batch_normalization_14[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_5 (DepthwiseCo (None, 14, 14, 240) 6240 activation_14[0][0]
__________________________________________________________________________________________________
batch_normalization_15 (BatchNo (None, 14, 14, 240) 960 depthwise_conv2d_5[0][0]
__________________________________________________________________________________________________
activation_15 (Activation) (None, 14, 14, 240) 0 batch_normalization_15[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_3 (Glo (None, 240) 0 activation_15[0][0]
__________________________________________________________________________________________________
dense_5 (Dense) (None, 60) 14460 global_average_pooling2d_3[0][0]
__________________________________________________________________________________________________
activation_16 (Activation) (None, 60) 0 dense_5[0][0]
__________________________________________________________________________________________________
dense_6 (Dense) (None, 240) 14640 activation_16[0][0]
__________________________________________________________________________________________________
activation_17 (Activation) (None, 240) 0 dense_6[0][0]
__________________________________________________________________________________________________
reshape_3 (Reshape) (None, 1, 1, 240) 0 activation_17[0][0]
__________________________________________________________________________________________________
multiply_3 (Multiply) (None, 14, 14, 240) 0 activation_15[0][0]
reshape_3[0][0]
__________________________________________________________________________________________________
conv2d_11 (Conv2D) (None, 14, 14, 40) 9640 multiply_3[0][0]
__________________________________________________________________________________________________
batch_normalization_16 (BatchNo (None, 14, 14, 40) 160 conv2d_11[0][0]
__________________________________________________________________________________________________
add_2 (Add) (None, 14, 14, 40) 0 batch_normalization_16[0][0]
batch_normalization_13[0][0]
__________________________________________________________________________________________________
conv2d_12 (Conv2D) (None, 14, 14, 240) 9840 add_2[0][0]
__________________________________________________________________________________________________
batch_normalization_17 (BatchNo (None, 14, 14, 240) 960 conv2d_12[0][0]
__________________________________________________________________________________________________
activation_18 (Activation) (None, 14, 14, 240) 0 batch_normalization_17[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_6 (DepthwiseCo (None, 14, 14, 240) 6240 activation_18[0][0]
__________________________________________________________________________________________________
batch_normalization_18 (BatchNo (None, 14, 14, 240) 960 depthwise_conv2d_6[0][0]
__________________________________________________________________________________________________
activation_19 (Activation) (None, 14, 14, 240) 0 batch_normalization_18[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_4 (Glo (None, 240) 0 activation_19[0][0]
__________________________________________________________________________________________________
dense_7 (Dense) (None, 60) 14460 global_average_pooling2d_4[0][0]
__________________________________________________________________________________________________
activation_20 (Activation) (None, 60) 0 dense_7[0][0]
__________________________________________________________________________________________________
dense_8 (Dense) (None, 240) 14640 activation_20[0][0]
__________________________________________________________________________________________________
activation_21 (Activation) (None, 240) 0 dense_8[0][0]
__________________________________________________________________________________________________
reshape_4 (Reshape) (None, 1, 1, 240) 0 activation_21[0][0]
__________________________________________________________________________________________________
multiply_4 (Multiply) (None, 14, 14, 240) 0 activation_19[0][0]
reshape_4[0][0]
__________________________________________________________________________________________________
conv2d_13 (Conv2D) (None, 14, 14, 40) 9640 multiply_4[0][0]
__________________________________________________________________________________________________
batch_normalization_19 (BatchNo (None, 14, 14, 40) 160 conv2d_13[0][0]
__________________________________________________________________________________________________
add_3 (Add) (None, 14, 14, 40) 0 batch_normalization_19[0][0]
add_2[0][0]
__________________________________________________________________________________________________
conv2d_14 (Conv2D) (None, 14, 14, 120) 4920 add_3[0][0]
__________________________________________________________________________________________________
batch_normalization_20 (BatchNo (None, 14, 14, 120) 480 conv2d_14[0][0]
__________________________________________________________________________________________________
activation_22 (Activation) (None, 14, 14, 120) 0 batch_normalization_20[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_7 (DepthwiseCo (None, 14, 14, 120) 3120 activation_22[0][0]
__________________________________________________________________________________________________
batch_normalization_21 (BatchNo (None, 14, 14, 120) 480 depthwise_conv2d_7[0][0]
__________________________________________________________________________________________________
activation_23 (Activation) (None, 14, 14, 120) 0 batch_normalization_21[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_5 (Glo (None, 120) 0 activation_23[0][0]
__________________________________________________________________________________________________
dense_9 (Dense) (None, 30) 3630 global_average_pooling2d_5[0][0]
__________________________________________________________________________________________________
activation_24 (Activation) (None, 30) 0 dense_9[0][0]
__________________________________________________________________________________________________
dense_10 (Dense) (None, 120) 3720 activation_24[0][0]
__________________________________________________________________________________________________
activation_25 (Activation) (None, 120) 0 dense_10[0][0]
__________________________________________________________________________________________________
reshape_5 (Reshape) (None, 1, 1, 120) 0 activation_25[0][0]
__________________________________________________________________________________________________
multiply_5 (Multiply) (None, 14, 14, 120) 0 activation_23[0][0]
reshape_5[0][0]
__________________________________________________________________________________________________
conv2d_15 (Conv2D) (None, 14, 14, 48) 5808 multiply_5[0][0]
__________________________________________________________________________________________________
batch_normalization_22 (BatchNo (None, 14, 14, 48) 192 conv2d_15[0][0]
__________________________________________________________________________________________________
conv2d_16 (Conv2D) (None, 14, 14, 144) 7056 batch_normalization_22[0][0]
__________________________________________________________________________________________________
batch_normalization_23 (BatchNo (None, 14, 14, 144) 576 conv2d_16[0][0]
__________________________________________________________________________________________________
activation_26 (Activation) (None, 14, 14, 144) 0 batch_normalization_23[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_8 (DepthwiseCo (None, 14, 14, 144) 3744 activation_26[0][0]
__________________________________________________________________________________________________
batch_normalization_24 (BatchNo (None, 14, 14, 144) 576 depthwise_conv2d_8[0][0]
__________________________________________________________________________________________________
activation_27 (Activation) (None, 14, 14, 144) 0 batch_normalization_24[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_6 (Glo (None, 144) 0 activation_27[0][0]
__________________________________________________________________________________________________
dense_11 (Dense) (None, 36) 5220 global_average_pooling2d_6[0][0]
__________________________________________________________________________________________________
activation_28 (Activation) (None, 36) 0 dense_11[0][0]
__________________________________________________________________________________________________
dense_12 (Dense) (None, 144) 5328 activation_28[0][0]
__________________________________________________________________________________________________
activation_29 (Activation) (None, 144) 0 dense_12[0][0]
__________________________________________________________________________________________________
reshape_6 (Reshape) (None, 1, 1, 144) 0 activation_29[0][0]
__________________________________________________________________________________________________
multiply_6 (Multiply) (None, 14, 14, 144) 0 activation_27[0][0]
reshape_6[0][0]
__________________________________________________________________________________________________
conv2d_17 (Conv2D) (None, 14, 14, 48) 6960 multiply_6[0][0]
__________________________________________________________________________________________________
batch_normalization_25 (BatchNo (None, 14, 14, 48) 192 conv2d_17[0][0]
__________________________________________________________________________________________________
add_4 (Add) (None, 14, 14, 48) 0 batch_normalization_25[0][0]
batch_normalization_22[0][0]
__________________________________________________________________________________________________
conv2d_18 (Conv2D) (None, 14, 14, 288) 14112 add_4[0][0]
__________________________________________________________________________________________________
batch_normalization_26 (BatchNo (None, 14, 14, 288) 1152 conv2d_18[0][0]
__________________________________________________________________________________________________
activation_30 (Activation) (None, 14, 14, 288) 0 batch_normalization_26[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_9 (DepthwiseCo (None, 7, 7, 288) 7488 activation_30[0][0]
__________________________________________________________________________________________________
batch_normalization_27 (BatchNo (None, 7, 7, 288) 1152 depthwise_conv2d_9[0][0]
__________________________________________________________________________________________________
activation_31 (Activation) (None, 7, 7, 288) 0 batch_normalization_27[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_7 (Glo (None, 288) 0 activation_31[0][0]
__________________________________________________________________________________________________
dense_13 (Dense) (None, 72) 20808 global_average_pooling2d_7[0][0]
__________________________________________________________________________________________________
activation_32 (Activation) (None, 72) 0 dense_13[0][0]
__________________________________________________________________________________________________
dense_14 (Dense) (None, 288) 21024 activation_32[0][0]
__________________________________________________________________________________________________
activation_33 (Activation) (None, 288) 0 dense_14[0][0]
__________________________________________________________________________________________________
reshape_7 (Reshape) (None, 1, 1, 288) 0 activation_33[0][0]
__________________________________________________________________________________________________
multiply_7 (Multiply) (None, 7, 7, 288) 0 activation_31[0][0]
reshape_7[0][0]
__________________________________________________________________________________________________
conv2d_19 (Conv2D) (None, 7, 7, 96) 27744 multiply_7[0][0]
__________________________________________________________________________________________________
batch_normalization_28 (BatchNo (None, 7, 7, 96) 384 conv2d_19[0][0]
__________________________________________________________________________________________________
conv2d_20 (Conv2D) (None, 7, 7, 576) 55872 batch_normalization_28[0][0]
__________________________________________________________________________________________________
batch_normalization_29 (BatchNo (None, 7, 7, 576) 2304 conv2d_20[0][0]
__________________________________________________________________________________________________
activation_34 (Activation) (None, 7, 7, 576) 0 batch_normalization_29[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_10 (DepthwiseC (None, 7, 7, 576) 14976 activation_34[0][0]
__________________________________________________________________________________________________
batch_normalization_30 (BatchNo (None, 7, 7, 576) 2304 depthwise_conv2d_10[0][0]
__________________________________________________________________________________________________
activation_35 (Activation) (None, 7, 7, 576) 0 batch_normalization_30[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_8 (Glo (None, 576) 0 activation_35[0][0]
__________________________________________________________________________________________________
dense_15 (Dense) (None, 144) 83088 global_average_pooling2d_8[0][0]
__________________________________________________________________________________________________
activation_36 (Activation) (None, 144) 0 dense_15[0][0]
__________________________________________________________________________________________________
dense_16 (Dense) (None, 576) 83520 activation_36[0][0]
__________________________________________________________________________________________________
activation_37 (Activation) (None, 576) 0 dense_16[0][0]
__________________________________________________________________________________________________
reshape_8 (Reshape) (None, 1, 1, 576) 0 activation_37[0][0]
__________________________________________________________________________________________________
multiply_8 (Multiply) (None, 7, 7, 576) 0 activation_35[0][0]
reshape_8[0][0]
__________________________________________________________________________________________________
conv2d_21 (Conv2D) (None, 7, 7, 96) 55392 multiply_8[0][0]
__________________________________________________________________________________________________
batch_normalization_31 (BatchNo (None, 7, 7, 96) 384 conv2d_21[0][0]
__________________________________________________________________________________________________
add_5 (Add) (None, 7, 7, 96) 0 batch_normalization_31[0][0]
batch_normalization_28[0][0]
__________________________________________________________________________________________________
conv2d_22 (Conv2D) (None, 7, 7, 576) 55872 add_5[0][0]
__________________________________________________________________________________________________
batch_normalization_32 (BatchNo (None, 7, 7, 576) 2304 conv2d_22[0][0]
__________________________________________________________________________________________________
activation_38 (Activation) (None, 7, 7, 576) 0 batch_normalization_32[0][0]
__________________________________________________________________________________________________
depthwise_conv2d_11 (DepthwiseC (None, 7, 7, 576) 14976 activation_38[0][0]
__________________________________________________________________________________________________
batch_normalization_33 (BatchNo (None, 7, 7, 576) 2304 depthwise_conv2d_11[0][0]
__________________________________________________________________________________________________
activation_39 (Activation) (None, 7, 7, 576) 0 batch_normalization_33[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_9 (Glo (None, 576) 0 activation_39[0][0]
__________________________________________________________________________________________________
dense_17 (Dense) (None, 144) 83088 global_average_pooling2d_9[0][0]
__________________________________________________________________________________________________
activation_40 (Activation) (None, 144) 0 dense_17[0][0]
__________________________________________________________________________________________________
dense_18 (Dense) (None, 576) 83520 activation_40[0][0]
__________________________________________________________________________________________________
activation_41 (Activation) (None, 576) 0 dense_18[0][0]
__________________________________________________________________________________________________
reshape_9 (Reshape) (None, 1, 1, 576) 0 activation_41[0][0]
__________________________________________________________________________________________________
multiply_9 (Multiply) (None, 7, 7, 576) 0 activation_39[0][0]
reshape_9[0][0]
__________________________________________________________________________________________________
conv2d_23 (Conv2D) (None, 7, 7, 96) 55392 multiply_9[0][0]
__________________________________________________________________________________________________
batch_normalization_34 (BatchNo (None, 7, 7, 96) 384 conv2d_23[0][0]
__________________________________________________________________________________________________
add_6 (Add) (None, 7, 7, 96) 0 batch_normalization_34[0][0]
add_5[0][0]
__________________________________________________________________________________________________
conv2d_24 (Conv2D) (None, 7, 7, 576) 55872 add_6[0][0]
__________________________________________________________________________________________________
batch_normalization_35 (BatchNo (None, 7, 7, 576) 2304 conv2d_24[0][0]
__________________________________________________________________________________________________
activation_42 (Activation) (None, 7, 7, 576) 0 batch_normalization_35[0][0]
__________________________________________________________________________________________________
global_average_pooling2d_10 (Gl (None, 576) 0 activation_42[0][0]
__________________________________________________________________________________________________
reshape_10 (Reshape) (None, 1, 1, 576) 0 global_average_pooling2d_10[0][0]
__________________________________________________________________________________________________
conv2d_25 (Conv2D) (None, 1, 1, 1024) 590848 reshape_10[0][0]
__________________________________________________________________________________________________
activation_43 (Activation) (None, 1, 1, 1024) 0 conv2d_25[0][0]
__________________________________________________________________________________________________
conv2d_26 (Conv2D) (None, 1, 1, 1000) 1025000 activation_43[0][0]
__________________________________________________________________________________________________
reshape_11 (Reshape) (None, 1000) 0 conv2d_26[0][0]
==================================================================================================
Total params: 2,555,742
Trainable params: 2,543,598
Non-trainable params: 12,144
__________________________________________________________________________________________________
Process finished with exit code 0
'''
