LoRA fine-tuning case study: Qwen2.5-1.5B for financial sentiment classification - View it on GitHub