paired-based method

[Zilong-Zhang003]

Very good work, but I didn't understand the meaning of the paired-based method. I'm not clear on what is being input to the model and where exactly it corresponds in the code. Looking forward to your reply.

[Zilong-Zhang003]

Paired-based inference. How many images are input to the model each time? Is it one input image at a time, optimized to match the target image, or is the input both the original image and a reference color image?