DPS2022 / diffusion-posterior-sampling

Official pytorch repository for "Diffusion Posterior Sampling for General Noisy Inverse Problems"
https://dps2022.github.io/diffusion-posterior-sampling-page/
402 stars 45 forks source link

Issue with phase retrieval and Poisson noise #9

Open TaliLeiba opened 12 months ago

TaliLeiba commented 12 months ago

Hello, I am trying to reproduce phase retrieval results with Poisson noise. I have tried different rates for the noise but I never got a decent result, while results for gaussian noise were good. are the values needed for phase retrieval with poison noise different from those who needed for PR with gaussian noise?

Thanks!

UnknownJenny commented 4 weeks ago

hello, I am trying to reproduce phase retrieval results with gaussian noise but still can't get good results. Have you change anything about the phase_retrival_config or the diffusion config for your good gaussian result? I was using ps for the cond_method, and scale 1.0, literally just the initial config the readme gives, but the result was too bad

UnknownJenny commented 4 weeks ago

Oh I just

hello, I am trying to reproduce phase retrieval results with gaussian noise but still can't get good results. Have you change anything about the phase_retrival_config or the diffusion config for your good gaussian result? I was using ps for the cond_method, and scale 1.0, literally just the initial config the readme gives, but the result was too bad

Oh I just figured it out. Turns out the result is randomly broke, somtimes good sometimes not.