Hands-On Deep Learning Algorithms with Python

Defining the decoder

Now, we define our decoder; in the decoder, we perform the deconvolution operation with three layers, that is, we upsample the encodings created by the encoder and reconstruct the original image.

Define the first convolutional layer followed by upsampling:

x = Conv2D(8, (3, 3), activation='relu', padding='same')(encoder)x = UpSampling2D((2, 2))(x)

Define the second convolutional layer with upsampling:

x = Conv2D(8, (3, 3), activation='relu', padding='same')(x)x = UpSampling2D((2, 2))(x)

Define the final convolutional layer with upsampling:

x = Conv2D(16, (3, 3), activation='relu')(x)x = UpSampling2D((2, 2))(x)decoded = Conv2D(1, (3, 3), activation='sigmoid', padding='same')(x)