search

wyhuai / DDNM

[ICLR 2023 Oral] Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Model
MIT License
1.16k stars 83 forks source link

What is the sampling configuration for DIV2K 0810.png in Figure 1(a)? #17

Closed kevin-on closed 1 year ago

kevin-on commented 1 year ago

Hello, I hope this message finds you well. First of all, I wanted to express my admiration for the outstanding work you have done on your repository.

I'm currently working on a project that requires the use of the DIV2K 0810.png image shown in Figure 1(a) in the paper. However, I couldn't find any information about the sampling configuration used to generate this image. If it's not too much trouble, I would like to kindly request information on the sampling configuration used to generate this image.

Thank you very much for your help!

wyhuai commented 1 year ago

Hi, actually I use a guided diffusion model pre-trained on ImageNet 512x512 (not yet implemented in this repository), with classifier guidance, T=1000, time-travel length 10, repeat 3.

kevin-on commented 1 year ago

Hello, thank you for your response.

I have a couple of follow-up questions. (1) What is the class label that is used to guide the image? (2) Is the resolution of the image shown in Figure1(a) is 512x512? I ask because the original image in the DIV2K dataset appears to have a larger resolution.

Thank you in advance for your help.

wyhuai commented 1 year ago

Hi, for the label, you can search "squirrel" on this link, and get 335: 'fox squirrel, eastern fox squirrel, Sciurus niger'.

The image in Fig.1(a) is 512x512. However, it is possible to solve higher resolution using a DDNM of only 256x256 resolution. We will submit a corresponding new work to arXiv soon.

kevin-on commented 1 year ago

Thank you so much for your kind assistance. I look forward to seeing your future work.