Open jojupiter opened 1 year ago
you said in your paper that you only use synthetic data to train your model. However when I see your dataset : div2K , flickr2K, they are photographs of real images
Is real-ESRGAN trained only with synthetic images?
"Synthetic" means that LR images are generated by self-defined degradations.
you said in your paper that you only use synthetic data to train your model. However when I see your dataset : div2K , flickr2K, they are photographs of real images
Is real-ESRGAN trained only with synthetic images?