Hi, I find the checkpoints/models released in this repo is a bit tricky for me since it is not mentioned in the paper.
In your paper, you apply the trained model to HSI denoising under different noise setting. By default, I would assume you use a single trained model to tackle all different noise setting.
But in your repo, you release two models, i.e. HSID-CNN_Realdata_iter_1000000 and HSID-CNN_noiselevel100_iter_600000.
So I was wondering, is that means you train multiple models to deal with each case respectively?
What's the model "HSID-CNN_Realdata_iter_1000000" resort to? How is that trained?
Thank you for your attention and looking forward to your reply.