andreas128 / RePaint

Official PyTorch Code and Models of "RePaint: Inpainting using Denoising Diffusion Probabilistic Models", CVPR 2022

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Resample Params

Siyeong-Lee opened this issue · comments

Hi, authors!

Thank you for releasing your work!
Could you tell me which parameters are related to '4.2 Resampling' in configs?
Could you please explain in detail the relevant parameters?


schedule_jump_params:
t_T: 250 # I think this parameter might mean the number of diffusion steps
n_sample: 1 # What does this parameter mean?
jump_length: 10 # this parameter might mean the jump length.
jump_n_sample: 10 # this parameter might mean the number of resamplings by r.

I am also interested in more details on the resample part. Thank you.

I am wondering why the sampling starts from 250.