search

swz30 / MPRNet

[CVPR 2021] Multi-Stage Progressive Image Restoration. SOTA results for Image deblurring, deraining, and denoising.
Other
1.16k stars 188 forks source link

Deraining reproduction result seems a little different from report result on paper #100

Open peylnog opened 2 years ago

peylnog commented 2 years ago

Because the code using the torch.backends.cudnn.benchmark=True , I know it cannot completely repetition. Then , I reproduced it on my 8x2080TI without any code change, and here's result I got.

result

May I report this result ( reproduce without any code change) on my paper ? It is very important for my work, since I need to improve the performance based on your code and model.

If there are other factors I haven't considered, please tell me, and I'll reproduce again.

Looking forward to your reply!

Have a good day~

hezewy commented 2 years ago

I also had the same problem.

LunZeHu commented 1 year ago

I also had the same problem. I reproduced it on my RTX3060 without any code change, and the result seem different from report result on paper. I know it cannot completely repetition. May I report My result ( reproduce without any code change) on my paper? It is very important for my work. If there are other factors I haven't considered, please tell me, and I'll reproduce again. Looking forward to your reply! Have a good day~