Autoencoder. Midpoint.

KNIME Extensions Deep Learning
#1

Hi there,

I have been analyzing this kind of workflows relating to Autoencoders:

However, I am trying to use the Autoencoder as a dimensional reduction tool. That is to say, train the Autoencoder and then get the results in the mid point at the end of the encoder. I have done it in TensorFlow but I would like to do it with Knime. However, How could I get the model of the “encoder” if the model is trained as a whole?

Thanks in advance

Image denoiser with autoencoder
#2

Upsss I got it by myself!!! The Keras Network Executor lets you chose which layer to have in the output. So nice!!

4 Likes