Linked Upstream Epic:

• https://github.com/instructlab/training/issues/362 

Goal:

• The `instructlab/training` library has been successful shipped in RHEL AI 1.y releases. For the purpose of 1.y releases, the library wasn't supposed to be consumed directly- it was always consumed by `ilab` directly, and could be plumbed as necessary.
• For 2.y and future releases, the library must be consumable by AI Engineering practitioners as an SDK, presumably via Jupyter Notebooks.
• This necessitates a thorough overview of the library usability, understandability, tests, and documentation.

Acceptance Criteria:

The Acceptance Criteria provides a definition of scope and the expected outcomes - from a users point of view - defines the value proposition

• The `instructlab.training` repo should be usable in a jupyter notebook environment, easily configurable with sensible defaults, that is independent of the `ilab` workflow as much as possible.
• It should implement SFT and RLHF techniques necessary for success in the ilab workflow.
• It should be modular / composable.