search

IceClear / StableSR

[IJCV2024] Exploiting Diffusion Prior for Real-World Image Super-Resolution
https://iceclear.github.io/projects/stablesr/
Other
2.23k stars 146 forks source link

SNR in the encoder #86

Open LuigiSigillo opened 1 year ago

LuigiSigillo commented 1 year ago

In the paper you said that

our proposed encoder is designed to offer comparatively strong conditions to the diffusion model within the range where the SNR hits 5e−2.

Where exactly you implemented this on the code? I have not found anything on checking the SNR before "strong conditioning" the diffusion model. Thank you.

IceClear commented 1 year ago

Hi. SNR is used as a metric to analyze our time-aware encoder, not a conditioned input. So we did not add the measuring code inside the repo.