chore!: Rename model_path to model in the Llama.cpp integration

[ZanSara]

Related Issues:

• part of Rename model_name and model_name_or_path parameters as model in all integrations #227
• the PR for the card is still open here: Add llama.cpp Integration haystack-integrations#118 Changes were suggested accordingly

Proposed Changes:

• Rename model_path to model in the embedders of the Llama.cpp integration

How did you test it?

Local tests run, CI

Note for the reviewer

This integration is a bit different from the others because it only expects a model path, so the name might be justified. However, for consistency I preferred to standardize this one as well. Feel free to object this change if we prefer to leave the parameter's name unchanged instead.