RHEL AI 1.5.0, when release, will only support Inferencing on multi Gaudi 3 AI Accelerators.

INTEL is working on patches to bring support for Fine Tuning. And more precisely, Training.

INTEL has started their internal process to raise those PRs into the corresponding github repos (https://github.com/instructlab/instructlab and https://github.com/instructlab/training)

We'd like for RHEL AI pick up those INTEL patches as part of an upcoming RHEL AI 1.5 z-stream release when those have been merged.

Upstream PR#552