How to train KBNet in raw image denoising task?

[e96031413]

I hope this message finds you well. I recently came across your project KBNet and I'm quite interested in using it for raw image denoising task. However, I couldn't find detailed instructions on how to train the model in the documentation. Therefore, I would greatly appreciate it if you could provide some guidance or point me in the right direction.

Specifically, I have the following questions regarding the training process:

1. What are the recommended steps to prepare the training data for KBNet?
2. Are there any specific data formats or preprocessing requirements that I need to follow?
3. Could you please provide an example command or configuration for training KBNet on SIDD dataset?

I understand that you may be busy, but any help or insights you can provide would be immensely valuable.

Thank you for your time and attention. I look forward to hearing from you.