[SANA-WM] Request: Release the pre-processed ViPE results for the full training dataset

Hi SANA-WM team,

Thank you for your impressive work and for open-sourcing the project.

I am currently working with the SANA-WM training pipeline. However, I’ve found that the ViPE processing is extremely time-consuming, especially when dealing with the full training dataset. This step has become a significant bottleneck in terms of both time and computational resources.

To help the community speed up research and ensure reproducibility, would it be possible for you to release the pre-processed ViPE/camera results for the full training set? Having access to these pre-processed files would save users a lot of redundant computation and allow us to focus more on the model training and evaluation.

Thank you for your time and consideration! Best regards.