Query: Diffusion models vs GANs

[KomputerMaster64]

@prafullasd @unixpickle Thank you for your work. I have a query, while catching up with the research in diffusion models, I stumbled upon one paper which asks for the comeback of GANs for they require less parameters and hence scaling models like BigGAN could be a good approach. Which approach should one now adopt to help diffusion models beat GANs again? reference: StudioGAN (benchmark) https://arxiv.org/abs/2206.09479