dangeng / visual_anagrams

Code for the paper "Visual Anagrams: Generating Multi-View Optical Illusions with Diffusion Models"
MIT License
870 stars 81 forks source link

Deepfloyd alternative? #9

Closed malik0 closed 10 months ago

malik0 commented 10 months ago

Hi, Is there an alternative we can use instead of DeepFloyd? thanks

dangeng commented 10 months ago

Hey! The method requires a pixel diffusion model, and I'm unfortunately not aware of any other very very good pixel diffusion models that are public. There's Imagen, but it's not public. You could also try using Stable Diffusion, which this colab notebook does, but I think the quality will be a bit worse because it's doing diffusion in latent space (see the paper for details). If you are aware of any really really good public pixel-diffusion models, I would love to know!