search

IceClear / StableSR

[IJCV2024] Exploiting Diffusion Prior for Real-World Image Super-Resolution
https://iceclear.github.io/projects/stablesr/
Other
2.23k stars 146 forks source link

Training and Test datasets #87

Open peylnog opened 1 year ago

peylnog commented 1 year ago

Thank you so much for this excellent work. Can you open source the training set and test datasets used in the paper.

IceClear commented 1 year ago

Hi. For training sets, you can refer to here For test sets, you may refer to the link in readme.

peylnog commented 1 year ago

Thanks for your reply. I have downloaded these test sets and I am curious why you crop them to 128?

IceClear commented 1 year ago

The original resolution of the test set can reach 2K or larger. The test time can be huge.

peylnog commented 1 year ago

Thank you. By the way. Are the RealSR, DRealSR, and DPED-iphone test datasets evaluated after retraining on the corresponding original training datasets, or are they directly tested on these datasets after training on datasets such as DIV2K as mentioned in the paper to evaluate generalization of proposed method?

IceClear commented 1 year ago

We do not finetune on real-world pair datasets. For details you may refer to our paper.

Juan-Chen45 commented 2 months ago

Thanks for the great work, I wonder if it is possible to open-source the scripts used for pre-processing testsets (adding degradation etc.)? Thank you. @IceClear

ZH3nGsQ commented 1 month ago

Thanks for your reply. I have downloaded these test sets and I am curious why you crop them to 128?

Could you plz tell me how to download those test datasets? I didn't find link in readme to download test datasets.