Keras Autoencoders - How to visualize the values of the hidden layer?

Question

I am trying to do a project where the values of hidden layers play a pivotal role. I am trying to use a sample autoencoder from this tutorial, https://blog.keras.io/building-autoencoders-in-keras.html I am able to do the gradient descent and it is also converging, but I am not sure how to print the values of the hidden layer. When I use print state on the model.outputs, I get tf.Tensor 'add:0' shape=(?, 30) dtype=float32, where 30 is number of nodes in the hidden layer. Can anyone help? Thanks.

Thomas Pinetz · Accepted Answer · 2017-03-16 08:45:50Z

1

This has to be done using Keras functions as you can read here: (https://keras.io/getting-started/faq/#how-can-i-obtain-the-output-of-an-intermediate-layer).

In essence you build a function like this:

import keras.backend as K
output_func = K.function([model.layers[0].input, K.learning_phase()],
                     [model.layers[1].output])
intermediate_output = output_func([data, False])

answered Mar 16, 2017 at 8:45

Thomas Pinetz

7,1882 gold badges31 silver badges48 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Keras Autoencoders - How to visualize the values of the hidden layer?

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related