Fine-tuning Qwen 2.5 3B for RBI Regulations: Achieving 8x Performance with Smart Data Augmentation

A deep dive into fine-tuning LLMs for domain-specific Q&A using LoRA, Unsloth, and data augmentation strategies