Experiment manager with feed forward architecture

I'm wondering if it is possible to use the Experiment Manger App to run an experiment with a feed-forward neural net. There are no documentation about it, and I have found no tutorial to help me use the app in this way.

I was doing several test by changing number or neurons, layers, activation functions,...but if it is possible to set up an experiment on that app and test several configuration and architectures will be better.

Anyone knows if it is possible and if it's so, how can I set up every part of the experiment?

Thank you in advance

Answer 1

[Khoo Yit Phang]

Unfortunately, training feed-forward neural net currently isn't supported in Experiment Manager (as of R2020b), it only supports training deep networks (e.g., trainNetwork).