Aloha, I am using the 7.2.0 beta2 build of 16/12 on Win 10. SQLite. Run the standard facial reco (took about a day). Started running the YOLO based one about 4 days ago. Completed about 30% of the images and still running. Yesterday, I lost the digikam UI. The process is still running, opening new images and attempting to identify faces. thumbnails-digikam.db is constantly modified. Found out that I can now open a new digikam session (new process). This one will have UI. I closed it to avoid conflicts. The first digikam process still runs and I would like to let it run its course (probably around a week more). Unless there is a way to stop the YOLO reco process and then run YOLO on the files it did not analyse yet (as opposed to starting from the beginning). I would be happy to provide additional info. Take Care; --Haim
How lost the UI? Then digiKam probably crashed. You don't have to do facial recognition in one go. You can set that images that have already been scanned are skipped. Maik
It doesn't make sense to let the processes continue if digiKam actually crashed. I am really surprised that face detection should still run in the background. Maik
Why are you still using the December 16th version? I think it won't include the fix for PSD files yet because you had PSD files that caused a crash. Maik
Thanks. The 16/12 build includes the PSD fix. I scanned all the files w/o YOLO. I then want to get YOLO working on all the files. I.e., rescan and merge. If I stop and then start the YOLO scan again it assumes all the files were scanned (which is true as they were scanned w/o YOLO). I am not aware of a way to get YOLO starting on the files YOLO did not scan. Thanks; --Haim
The strange crash with the face detection has been fixed in digiKam-7.2.0-RC, I close the bug. Maik