I've uploaded a video and reviewed it's automatic captions, which worked great really. In some places I had to manually correct it (I uploaded a short clip from The Neverending story, and the model had difficulty distinguishing some of the names, such as Atreyu and Bastian).
After that, towards the end of the clip several sentences were missed by the model. When I tried adding them manually all I got was the (not extremely informative) error message "Couldn't add transcript line"). My added line didn't overlap with other caption timestamps.
Any ideas why this is happening?