I'm not sure if someone's written a tutorial yet for the generation method I'm using, which is CLIP Guided Diffusion, but I believe @images_ai has a pinned tweet with a tutorial on how to use VQGAN+CLIP! It's a different method, so it gives different results,