README.md

metadata

title: FinanceBench-LLM Financial QA
emoji: 📊
colorFrom: green
colorTo: gray
sdk: gradio
sdk_version: 5.20.0
python_version: '3.12'
app_file: app.py
pinned: true
license: mit
tags:
  - finance
  - nvidia
  - lora
  - llm
  - sec-filings
  - financial-qa

FinanceBench-LLM: Domain-Adapted Financial QA

Built with NVIDIA NIM, NeMo Customizer (LoRA fine-tuning), and evaluated with LLM-as-a-Judge on the FinanceBench dataset.

Results

Model	Exact Match	F1 Score	ELO Rating
Base (Llama-3.1-8B)	0.23	0.41	835
+ ICL (5-shot)	0.34	0.56	1023
+ LoRA Fine-tuned	0.52	0.71	1142

+126% Exact Match improvement through LoRA fine-tuning on financial QA data.

Links

• GitHub Repository
• NVIDIA DLI Course