Merging multiple SLEAP projects while keeping only labelled frames (not full videos)

[2919387]

I’m working with a large dataset (60 videos, 10 minutes each, 30 fps, 2.7 k resolution).
If I load all videos into a single SLEAP project, the GUI crashes or slows down a lot, so I’ve been labelling subsets of videos in separate projects (e.g. project_1.slp, project_2.slp, etc.).

Ultimately, I’d like to merge these projects so that I can train a single model on all of my labelled frames.

Is it possible to merge the labels while keeping them attached to the extracted labelled frames only, rather than referencing the full video files? In other words, can I carry across only the frames that were labelled (and their labels) when merging, to avoid re-loading all of the original videos and crashing SLEAP?

Thanks a lot!