You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Your LR image seems to have strong JPG compression artifacts. The published models have not been trained to remove these artifacts. Instead high-quality HR images were transformed to LR images via bicubic downscaling (see DIV2K dataset) and used for model training. See also #19 for a related discussion. If you want a model to deal with JPG compression artifacts you'd need to create your own dataset for model training.
thank you for giving information, I have only low-quality images I don't have any high-quality image but, I saw your code their you sending DIV2K_high_train Images but I don't have I want to train my images my low to high-quality resolution please can you help me.
Still my images its coming low quality only please can you provide me Generator code I have to check
The text was updated successfully, but these errors were encountered: