search

zsyOAOA / ResShift

ResShift: Efficient Diffusion Model for Image Super-resolution by Residual Shifting (NeurIPS@2023 Spotlight, TPAMI@2024)
Other
944 stars 50 forks source link

How to test ResShift in RealSR x4 dataset? #18

Closed cswry closed 1 year ago

cswry commented 1 year ago

The LR resolution and GT resolution are the same in the RealSRx4 [1] Dataset.

It doesn't work whenI just set params "--scale" from 4 to 1 in inference.py.

Maybe I should downsample LR first?

Looking forward your reply!

Thank you!

[1]. Jianrui Cai, Hui Zeng, Hongwei Yong, Zisheng Cao, and Lei Zhang. Toward real-world single image super-resolution: A new benchmark and a new model. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), pages 3086–3095, 2019.

zsyOAOA commented 1 year ago

You should need to downsample the LR firstly.

Note that RealSR dataset have different versions. In version3, the LR/HR images have different resolutions. Please carefully read the introduction for this dataset on the website. @cswry

cswry commented 1 year ago

You should need to downsample the LR firstly.

Note that RealSR dataset have different versions. In version3, the LR/HR images have different resolutions. Please carefully read the introduction for this dataset on the website. @cswry

Oh!!! Thanks!!!