CNN网络简单优化，以VGG16为例

优化常用经验方法

1、调低学习率（或按迭代次数衰减）
2、调整参数的初始化方法
3、调整输入数据的标准化方法
4、修改Loss函数
5、增加正则化
6、使用BN/GN层（中间层数据的标准化）
7、使用dropout

VGG16网络结构

以下代码仅对第一个Conv2D进行了调整，增加了L2正则化，BN层与Dropout层，后面的的Conv2D均可进行相应调整

num_classes = 10
weight_decay = 0.000

model = keras.models.Sequential()

# 优化 增加L2正则化
model.add(keras.layers.Conv2D(64, (3, 3), padding='same', kernel_regularizer=keras.regularizers.l2(weight_decay)))
model.add(keras.layers.Activation('relu'))
# 优化 添加BN层和Dropout
model.add(keras.layers.BatchNormalization())
model.add(keras.layers.Dropout(0.3))

model.add(keras.layers.Conv2D(64, (3, 3), padding='same', activation='relu'))
model.add(keras.layers.MaxPooling2D(pool_size=(2, 2), strides=2))

model.add(keras.layers.Conv2D(128, (3, 3), padding='same', activation='relu'))
model.add(keras.layers.Conv2D(128, (3, 3), padding='same', activation='relu'))
model.add(keras.layers.MaxPooling2D(pool_size=(2, 2), strides=2))

model.add(keras.layers.Conv2D(256, (3, 3), padding='same', activation='relu'))
model.add(keras.layers.Conv2D(256, (3, 3), padding='same', activation='relu'))
model.add(keras.layers.MaxPooling2D(pool_size=(2, 2), strides=2))

model.add(keras.layers.Conv2D(512, (3, 3), padding='same', activation='relu'))
model.add(keras.layers.Conv2D(512, (3, 3), padding='same', activation='relu'))
model.add(keras.layers.MaxPooling2D(pool_size=(2, 2), strides=2))

model.add(keras.layers.Conv2D(512, (3, 3), padding='same', activation='relu'))
model.add(keras.layers.Conv2D(512, (3, 3), padding='same', activation='relu'))
model.add(keras.layers.MaxPooling2D(pool_size=(2, 2), strides=2))

model.add(keras.layers.Flatten())
model.add(keras.layers.Dense(256, activation='relu'))  # VGG16为4096
model.add(keras.layers.Dense(128, activation='relu'))  # VGG16为4096
model.add(keras.layers.Dense(num_classes, activation='softmax'))  # VGG16为1000