Hi, I'm trying this implementation using my custom dataset. Is it possible to train without the ground truth images? If no, is there some "easy way" to generate them? Thank you.
Hi,
I'm trying this implementation using my custom dataset.
Is it possible to train without the ground truth images?
If no, is there some "easy way" to generate them?
Thank you.