Deep Learning By Example

Building the model

In this implementation, we will be using more layers in the encoder and decoder part, and the reason for this is the new complexity that we have added to the input.

The next model is exactly the same as the previous CAE but with extra layers that will help us to reconstruct a noise-free image from a noisy one.

So let's go ahead and build this architecture:

learning_rate = 0.001# Define the placeholder variable sfor the input and target valuesinputs_values = tf.placeholder(tf.float32, (None, 28, 28, 1), name='inputs_values')targets_values = tf.placeholder(tf.float32, (None, 28, 28, 1), name='targets_values')# Defining the Encoder part of the netowrk# Defining the first convolution layer in the encoder parrt# The output ...