Several StyleGAN2 models to add

[torridgristle]

https://github.com/jeffheaton?tab=repositories&q=pretrained&type=&language=
Minecraft 1024px, Fish 256px, 1970s Sci-Fi 1024px, Christmas 256px

Judging by the fakes000000.png images from each, they weren't using transfer learning.

https://github.com/joshchen984/pretrained-stylegan2-hedgehogs
Hedgehogs 256px

https://github.com/dobrosketchkun/NeuralKuvshinov
https://drive.google.com/drive/folders/1T6BNlyPpvLsxXI-gkCepR7S_nFcSi63o (larger number = trained longer)
kuvshinov_ilya instagram artwork, transfer trained from stylegan2-ffhq-config-f.pkl

[justinpinkney]

Oh wow! Thanks for all the tips ☺ Will get on to adding them shortly!

[johndpope]

Hi Jusin, I had asked elsewhere - but maybe you didn't catch - with your toonifaction api / do you have any idea on how to speed up process? openapi gives an instant tooned photo - and I can't replicate this - although presumably they collapse the training to a single batch....

[justinpinkney]

@johndpope The obvious way to speed up inference is to run it on a GPU. I could do this and then it would take <1 second (probably network transfer would be the slowest bit). It's just a question of volume/price to make it worth running on GPUs (which are obviously much more expensive than sticking to CPU). It's also a question of actually having the time to do it, but I already run some stuff on autoscaling GPUs, so it's totally doable.