How to customize the "model function" in the deep learning toolbox to add learnable parameters?

It seems that at present, you can only add learnable parameters in the form of "Define Custom Deep Learning Layer with Learnable Parameters", but this method is very inflexible. If you change to "model function" to define the network, you don't know how to add learnable parameters. Parameters?

Answer 1

[Jack Xiao]

please see the demo of training gan.