search

The-Learning-And-Vision-Atelier-LAVA / DASR

[CVPR 2021] Unsupervised Degradation Representation Learning for Blind Super-Resolution
MIT License
388 stars 50 forks source link

blur kernels in Tab.3 #25

Closed sujyQ closed 3 years ago

sujyQ commented 3 years ago

Hi. I'm sorry to bothering you. Does this paper contain information about eigen values and rotate angles of the blur kernels in Tab.3 ? I can see just shapes.

LongguangWang commented 3 years ago

Hi @sujyQ, thanks for your interest in our work. The detailed settings of the blur kernels in Tab. 3 are shown below. image While re-producing the results in Tab. 3, we find a bug and fix it. We also notice that a wrong pre-trained model for anisotropic kernels is released by mistake and we have uploaded the correct one.

sujyQ commented 3 years ago

Thanks :)

Lewis0427 commented 2 years ago

util代码中random_anisotropic_gaussian_kernel里面您的更改是否有错误：theta = torch.rand(batch).cuda() / 180 math.pi 是否应是theta = torch.rand(batch).cuda() math.pi，180无需再加了

LongguangWang commented 2 years ago

@Lewis0427，感谢指出这一bug，我们已经对其进行了修复。