Support HuggingFace Transformers

[SamTheCoder777]

This PR introduces support for HuggingFace Transformers in Parsera by adding a new HuggingFaceModel. Users only need to install Transformers when they choose to use it, as the module is not loaded until Parsera is initialized with HuggingFaceModel as the model.

This PR introduces support for HuggingFace Transformers by adding a new class, ParseraHuggingFace. The class is designed for lazy loading, ensuring that users only install the transformers library if they choose to use it.

List of changes:

• Added: custom HuggingFaceModel in model.py
• Added: support for HuggingFaceModel in Parsera
• Updated: README with examples for using HuggingFaceModel with Parsera

- Added: ParseraHuggingFace class for HuggingFace integration.

[portoaj]

I think this a great feature but I think making a whole new ParseraHuggingFace is unnecessary unless there's something that I'm missing. I think it would be better if you could either make the ParseraHuggingFace class inherit _run, run, and arun from Parsera since those seem to be mostly untouched, or even better just add the HuggingFace support to the existing Parsera class potentially with some helper methods.

[SamTheCoder777]

I’ve removed the unnecessary ParseraHuggingFace class to avoid code duplication. Now, HuggingFaceModel integrates directly with Parsera as it's already a BaseChatModel instance.