Anyone have an intuition on how to upscale images produced from Stable-Diffusion?
I’ve been looking online and can only find this repo https://github.com/xinntao/Real-ESRGAN
If not, would like to explore this space with others!
Anyone have an intuition on how to upscale images produced from Stable-Diffusion?
I’ve been looking online and can only find this repo https://github.com/xinntao/Real-ESRGAN
If not, would like to explore this space with others!
Also check out the Cold Diffusion paper, which has some very interesting ideas in this direction.
2x upscaler diffusion model released by Katherine Crowson:
It’s a latent diffusion model trained specifically with the Stable Diffusion autoencoder latent space.