I am sorry bro, but that was a hell of a confusing read. I don't think I understood a thing of it (and I do research GANs). I will give the paper a try as it seems there are good ideas here.

Constructive feedback: you are taking too many concepts and things for granted. Figure 1 explains very little. What *.+ means? What are f and g funcitons? (are they functions?) What is e1 and e2? On Figure 2, why are there two latents z1 and z2? How is z2 generating a real image, wouldn't it be the other way around?

"Architecture details". I highly believe that a snippet of your TF/Keras/Pytorch code would explain way more than these two tables. Currently, I don't have a clear idea of what the notation is trying to convey. A fully connected layer with size 256x32x32 is what? I believe it is a Conv 1x1, but I can't be sure.

What 512 -> 256x256x32x32 mean?

Etc