Add support for LLaMA2

[danieldk]

Description

LLaMA2 support different number of query and key-value heads. Also add new models to test k-n and k-1 head configurations.

Types of change

Feature

Checklist

• I confirm that I have the right to submit this contribution under the project's MIT license.
• I ran the tests, and all new and existing tests passed.
• My changes don't require a change to the documentation, or if they do, I've added all required information.

[shadeMe]

Tests pass locally, but we need to increment the transformers pin in the CI.

EDIT: We'll also need to add a separate generator for the chat models as they need custom prompting.