swz30/MIRNetv2

Do images in MIRNetv2 need to be normalized to [0,1] or [-1, 1]?

Spacei567 opened this issue · 2 comments

Hi, Can you tell me in MIRNet v1 and MIRNet v2 wheather the images need to be normalized?

swz30 commented

Hi @Spacei567

The images are normalized to [0,1]

input_ = torch.from_numpy(img).float().div(255.).permute(2,0,1).unsqueeze(0).to(device)

Thank you so much! And you research is excellent!