[WIP] Feature/progress bar

[CarloLepelaars]

Fixes issue #20

• Adds progress bar to all text and image embedders.
• Tests for SentenceEncoder.
• Use perfplot for progress bar?
• Can we ensure fast NumPy vectorization while using a progress bar?

[koaning]

One thing that I'm also wondering is... Technically nothing is stopping you from doing something like this:

tfm = SentenceTransformer()
tfm.transform([t for t in tqdm(long_list_of_texts)])

It's not a huge ask from an end user to do this, and it might just be enough to add an FAQ on the docs for this.

Still, any performance number/chart on this would be grand to share there as well.

[CarloLepelaars]

Agree, if you are iterating one by one the user can easily add a progress bar and configure it as they like.

I'm not that familiar with mkdocs, but can have a look on integrating a perfplot in the docs.

[koaning]

There's a course on calmcode 😉

[CarloLepelaars]

Calmcode saves the day, as usual! 😁

[CarloLepelaars]

Closing this to avoid introducing unnecessary complexity and slow down these encoders.