Would you like to share your trained model and why your method is called 'Unsupervised'?

[ShihuaiXu]

thanks for your great work, I have run your code in my PC.
I just check the input of the model, it contains single RGB images and a .mat file which contains the depth information, and you use CNN to predict the depth of the input RGB image and use PSFLayer to refine it and count loss, so why it is called 'unsupervised', please help me.
Besides, I would like to use your trained model to test some images, could please share your trained model to me?
What's more, I wonder know what is the meaning of Focus in you paper, it is the parameter of the camera?