Fine-tune LLMs using Red Hat training-hub library with SFT, LoRA, and OSFT algorithms. Use when preparing JSONL datasets, running training jobs, configuring hardware, scaling to clusters, evaluating models, or deploying with vLLM.