Dear Participants,
I recently received feedback that "1_lightstage_subdataset_restructure.py removed all my old devPhase data and killed all my training jobs too". I immediately double-checked the starting kit and would like to share my findings with you.
The 1_lightstage_subdataset_restructure.py script is designed to move your devPhase data to the chaPhase folder, ultimately combining the separately released devPhase and chaPhase datasets. I can confirm that the original script is working as intended. However, please note that this script will move your devPhase files to the chaPhase folder. If you are still running training jobs using the devPhase data folder, I advise you to duplicate the directory containing the devPhase data so that you can continue your training process while preparing your chaPhase dataset.
I had assumed that changing folders would be a natural part of transitioning to the Challenge Phase, but I apologize for not explaining the details of the folder combining process more clearly in advance. I hope this explanation helps.
Best regards,
Youngkyoon Jang