How can I create a stack of autoencoders and a softmaxLayer from a Vector of autoencoders?

Question

Pablo Dias 2017-11-22

回答： DS 2018-4-6

I am trying to create an stack of autoencoders, from the code below:

stackedNet = stack(vector_autoencoders(:),softnet);

but I'm getting the following message: "Expected one output from a curly brace or dot indexing expression, but there were 2 results."

There are 2 autoencoders in my vector with proper dimensions. If I use the following code it works:

stackedNet = stack(vector_autoencoders(1),vector_autoencoders(2),softnet);

What am I doing wrong? Is there any other way to use stack, without enumerating every autoencoder?

Answer 1

DS 2018-4-6

Hi Pablo,

if vector_autoencoders is a cell array you can add your AE in a for loop by:

vector_autoencoders{end+1} = trainAutoencoder(...)

Afterwards, you can stack them as follows:

stackedNet = stack(vector_autoencoders{:,:},softnet);