Update _face_embedding.py

[mst-rajatmishra]

hello, i tried making some improvements in code, File Operations: In add_face, consider adding error handling for file operations (e.g., during saving/loading) to avoid crashes due to I/O issues. Face Detection: Ensure that the detect_faces method gracefully handles cases where no faces are detected, returning an empty list instead of raising an unhandled exception.

[w4hns1nn]

Hey rajat,

thanks for your contribution. I just had a brief look and looks solid to me. 💯 I'll check-it out and test it this weekend and then approve the pr.
👍

[w4hns1nn]

Hey Raj,

while i looked into it, i found some more bugs. For this reason I made some bigger changes:

1. A face embedding now can hold only one face (was probably confusing anyway before). It simplified the downstream implementation quite a bit. Of course i have included your change.
2. Removed upscaling as Standard setting for video swap
3. Added better progress bar for video swap

I'd be great if you could check the new implementation and if it works for you. I pushed everything to the dev branch.

Thanks!