search

wyhuai / DDNM

[ICLR 2023 Oral] Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Model
MIT License
1.16k stars 82 forks source link

MSR and HiR for latent diffusion model ? #21

Open lizhengwei1992 opened 1 year ago

lizhengwei1992 commented 1 year ago

Thanks for the great work ! I can not see codes example using HiR method . the solution of MSR and HiR, only work in pixel space, doesn't apply to latent space ,because the consistency can't be satisfied in latent space ? Consistency : Axˆ ≡ y, Realness : xˆ ∼ q(x),

wyhuai commented 1 year ago

Hi, the code of HiR is not updated yet in this repository (sorry, I'm too busy recently. But I think you may implement it by just changing a few lines of code following the paper).

For latent space, you can use x=D(z), where D is the decoder, and optimize the consistency constraints rather than use RND. This can work, but may not yield results as good as in the pixel domain with RND.