Code for the degradation pipeline

[yyNoBug]

According to your paper, you have created three synthetic datasets using a degradation pipeline. However, in your codebase, there is no code for the degradation pipeline. I've tried to replicate your dataset degradation strategy, but the description in your paper is not clear enough to provide all the details for the degradation pipeline.

Could you please provide your code for the degradation? This can be very helpful. Thanks!

[LaVieEnRoseSMZ]

Unofficial comments: I am trying to use this and adapt it from image degradation to video degradation https://github.com/cswry/OSEDiff/blob/main/dataloaders/realsr_dataset.py

[yyNoBug]

Thank you for the reply! Seems like the authors truly adapted an image degradation model for video degradation. Did you manage to reproduce the numbers in the paper with your degradation pipeline by the way?

[LaVieEnRoseSMZ]

I do not intend to reproduce the results, but I struggled with training due to the VAE decoder loss, which caused frequent out-of-memory issues.

[Teiboku]

https://github.com/NJU-PCALab/STAR/tree/main/utils_data Seems author has updated the degrade process, but it is very slow at my machine (A6000 PRO), how about your case?