Prepares datasets and training configurations for fine-tuning LLMs. Use when customizing a base model on domain-specific data using LoRA, QLoRA, or supervised fine-tuning.
