How to separately use the encoder of an Autoencoder in keras?

问题

I have trained the following autoencoder model:

input_img = Input(shape=(1, 32, 32))

x = Convolution2D(16, 3, 3, activation='relu', border_mode='same')(input_img)
x = MaxPooling2D((2, 2), border_mode='same')(x)
x = Convolution2D(8, 3, 3, activation='relu', border_mode='same')(x)
x = MaxPooling2D((2, 2), border_mode='same')(x)
x = Convolution2D(8, 3, 3, activation='relu', border_mode='same')(x)
encoded = MaxPooling2D((2, 2), border_mode='same')(x)


x = Convolution2D(8, 3, 3, activation='relu', border_mode='same')(encoded)
x = UpSampling2D((2, 2))(x)
x = Convolution2D(8, 3, 3, activation='relu', border_mode='same')(x)
x = UpSampling2D((2, 2))(x)
x = Convolution2D(16, 3, 3, activation='relu',border_mode='same')(x)
x = UpSampling2D((2, 2))(x)
decoded = Convolution2D(1, 3, 3, activation='sigmoid', border_mode='same')(x)

autoencoder = Model(input_img, decoded)
autoencoder.compile(optimizer='RMSprop', loss='binary_crossentropy')

autoencoder.fit(X_train, X_train,
            nb_epoch=1,
            batch_size=128,
            shuffle=True,
            validation_data=(X_test, X_test)]
            )

After training this autoencoder i want to use the trained encoder for a supervised line. How can i extract only the trained encoder part of this autoencoder model ?

回答1:

You can just create a model after training that only uses the encoder:

autoencoder = Model(input_img, encoded)

If you want to add further layers after the encoded portion, you can do that as well:

classifier = Dense(nb_classes, activation='softmax')(encoded)
model = Model(input_img, classifier)

来源：https://stackoverflow.com/questions/39551478/how-to-separately-use-the-encoder-of-an-autoencoder-in-keras

标签

python

neural-network

deep-learning

keras

autoencoder