>>67167374 (me)Another sample workflow, this one with 4x-animesharp that is then scaled down to 0.4 before encoded back to latent.
https://files.catbox.moe/iaz2rx.pngI'd like to do one where I don't scale it down at all before encoding to latent, just for completeness sake, but my computer cannot handle VAE encode on a 4096 x 4096 image