An end-to-end Financial Sentiment Analysis system built using Microsoft Phi-2 (2.7B) fine-tuned with QLoRA (4-bit quantization + LoRA adapters) for efficient, high-performance training on limited hardware. - View it on GitHub