Open hyungsupsup opened 2 months ago
Hello, thanks for the interest. I don't think there is going to be a problem in changing the nosie scheduling and predicitoning of diffusion models, since our method mainly touches using patch with positions for training to reduce the computational cost. We tried experiments with Stable Diffusion, which is DDPM based, and it worked well.
Hello, I was wondering if there would be any issues applying the Patch Diffusion approach to DDPM frameworks instead of EDM. Have you ever tried it? If not, do you think there would be any problems? Thank you.