VGG16Net

import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers

#模型
model = keras.Sequential()

# 1st Layer
model.add(layers.Conv2D(input_shape=(224,224,3),filters=64,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.Conv2D(filters=64,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.MaxPool2D(pool_size=(3,3),strides=(2,2),padding='same'))

# 2nd Layer
model.add(layers.Conv2D(filters=128,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.Conv2D(filters=128,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.MaxPool2D(pool_size=(3,3),strides=(2,2),padding='same'))

# 3rd Layer
model.add(layers.Conv2D(filters=256,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.Conv2D(filters=256,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.Conv2D(filters=256,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.MaxPool2D(pool_size=(3,3),strides=(2,2),padding='same'))

# 4th Layer
model.add(layers.Conv2D(filters=512,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.Conv2D(filters=512,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.Conv2D(filters=512,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.MaxPool2D(pool_size=(3,3),strides=(2,2),padding='same'))

# 5th Layer
model.add(layers.Conv2D(filters=512,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.Conv2D(filters=512,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.Conv2D(filters=512,kernel_size=[3,3],strides=(1,1),padding='same',activation='relu'))
model.add(layers.MaxPool2D(pool_size=(3,3),strides=(2,2),padding='same'))

# Passing it to a Fully Connected layer
model.add(layers.Flatten())

# 7th Fully Connected Layer
model.add(layers.Dense(4096,activation='relu'))
model.add(layers.Dropout(0.5))

# 7th Fully Connected Layer
model.add(layers.Dense(4096,activation='relu'))
model.add(layers.Dropout(0.5))

# Output Layer
model.add(layers.Dense(1000,activation='softmax'))

model.summary()

_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
conv2d_31 (Conv2D)           (None, 224, 224, 64)      1792      
_________________________________________________________________
conv2d_32 (Conv2D)           (None, 224, 224, 64)      36928     
_________________________________________________________________
max_pooling2d_13 (MaxPooling (None, 112, 112, 64)      0         
_________________________________________________________________
conv2d_33 (Conv2D)           (None, 112, 112, 128)     73856     
_________________________________________________________________
conv2d_34 (Conv2D)           (None, 112, 112, 128)     147584    
_________________________________________________________________
max_pooling2d_14 (MaxPooling (None, 56, 56, 128)       0         
_________________________________________________________________
conv2d_35 (Conv2D)           (None, 56, 56, 256)       295168    
_________________________________________________________________
conv2d_36 (Conv2D)           (None, 56, 56, 256)       590080    
_________________________________________________________________
conv2d_37 (Conv2D)           (None, 56, 56, 256)       590080    
_________________________________________________________________
max_pooling2d_15 (MaxPooling (None, 28, 28, 256)       0         
_________________________________________________________________
conv2d_38 (Conv2D)           (None, 28, 28, 512)       1180160   
_________________________________________________________________
conv2d_39 (Conv2D)           (None, 28, 28, 512)       2359808   
_________________________________________________________________
conv2d_40 (Conv2D)           (None, 28, 28, 512)       2359808   
_________________________________________________________________
max_pooling2d_16 (MaxPooling (None, 14, 14, 512)       0         
_________________________________________________________________
conv2d_41 (Conv2D)           (None, 14, 14, 512)       2359808   
_________________________________________________________________
conv2d_42 (Conv2D)           (None, 14, 14, 512)       2359808   
_________________________________________________________________
conv2d_43 (Conv2D)           (None, 14, 14, 512)       2359808   
_________________________________________________________________
max_pooling2d_17 (MaxPooling (None, 7, 7, 512)         0         
_________________________________________________________________
flatten_3 (Flatten)          (None, 25088)             0         
_________________________________________________________________
dense_11 (Dense)             (None, 4096)              102764544 
_________________________________________________________________
dropout_8 (Dropout)          (None, 4096)              0         
_________________________________________________________________
dense_12 (Dense)             (None, 4096)              16781312  
_________________________________________________________________
dropout_9 (Dropout)          (None, 4096)              0         
_________________________________________________________________
dense_13 (Dense)             (None, 1000)              4097000   
=================================================================
Total params: 138,357,544
Trainable params: 138,357,544
Non-trainable params: 0
_________________________________________________________________

版权声明：原创,转载请注明来源,否则律师函警告