The model I think is fairly self explanatory.

Due to the incredibly limited dataset, only really 4 pictures, I doubled it up with the black and white variant of the doujin but I think that at least with NovaFurry 13.5 it's a bit weak at a strength of 1, 2 seemed to work well.

https://vocaroo.com/1dCKgpymuk1M

Back to my devilish tricks.

Works well at a strength of 2

Previews made with Novafurry 13.5