If the video's resolution has a longer vertical length, the captions are generated based on that vertical format. However, for videos with 4K resolution or higher, motion captions may not be generated smoothly.
If the resolution is too high, try lowering it.
Also, check that the video is wider horizontally (landscape format), then try generating the captions again.
For reference, the size of animated captions can be adjusted in [General] > [Scale] or [Style] > [Font Size] options.