Jiaqi Ma

感谢~ 我从头训练了一遍width32的模型，目前在SIDD validation是39.8525/0.9591，在SIDD benchmark上是39.7/0.958。为什么大家都选择在validation上的结果？虽然论文很多都用了validation的值，但是validation的结果选择最佳模型，benchmark作为online server的report是不是更合理？

> @TerryYiDa you could simply use difference graph. Hello, could you explain more on it? What is 'difference graph'? Thanks!

Got it. Can you provide SSIM of your method on DND and SIDD?

> Thanks for your report, we will have a discuss and fix it ASAP The quick way to make ` --show-score-thr` effect in this version is to modify '/mmdet/engine/hooks/visualization_hook.py'. ```...

Ok, it should be easy to fix. I will check and pull a request this week.

> Yeah, it is truly effective, but following their setting, my training result only gets 49.55 dB PSNR in the validation dataset (lower than theirs 51.06 dB). And about theory...

Waiting for new links!