simple_net.py

def simple_net(shape):
    from keras.models import Model
    from keras.layers import Dense, Conv2D, MaxPooling2D, Flatten, Input, BatchNormalization, Dropout
    from keras.regularizers import l2
    # The layers of Convolution → Convolution → Pooling
    def ccp(filters, kernel_size=(3, 3), weight_decay=1e-4, dropout=0.2):
        def _ccp(x):
            x = Conv2D(filters, kernel_size, padding='same', kernel_regularizer=l2(weight_decay), activation='relu')(x)
            x = BatchNormalization()(x)
            x = Conv2D(filters, kernel_size, padding='same', kernel_regularizer=l2(weight_decay), activation='relu')(x)
            x = BatchNormalization()(x)
            x = MaxPooling2D(pool_size=(2, 2))(x)
            x = Dropout(dropout)(x)
            return x
        return _ccp

    inputs = Input(shape=train_X.shape[1:])
    
    x = ccp(32)(inputs)
    x = ccp(64, dropout=0.3)(x)
    x = ccp(128, dropout=0.4)(x)
    x = Flatten()(x)
    y = Dense(10, activation='softmax')(x)
    
    model = Model(inputs=inputs, outputs=y)
    model.compile(loss='categorical_crossentropy',
            optimizer='adam',
            metrics=['accuracy'])

    return model