ephemeral182 / snowformer Goto Github PK

Please provide relevant training code

Thanks for your good job. As title, Looking forward you release SRRS and Snow 100K pretrained model.

Hello, thanks for your work.

Do you know if randomly sampling two times has significant differences on the final performance?

Thanks.

Thanks for your amazing work!
Did you evaluate your model on other tasks, e.g. denoising, deblurring?

作者，您好。我看你们的所有去雪任务的工作中，在其他方法上（例如NAFNet和Restormer等）的结果都是你们自己重新训练的，但是我在重新训练过程中，NAFNet的结果远高于你论文中的结果。我想知道这是怎么回事呢？

Dear author, thanks for your excellent work.

A quick question is how to sample 2000 images from Snow 100K testing set?

2000 images from each subset (S/M/L) and compute average, or mix 3 subsets and then select 2000 images randomly?

Thanks for your reply.

Thank you for your wonderful work!
Could you please tell me how to get the Supplementary Materials that you mention in the paper? Thanks a lot.