FinGPT

FinGPT: Open-Source Financial Large Language Models

What is FinGPT? FinGPT is an open-source financial large language model (LLM) designed to democratize access to financial data and analysis. Unlike proprietary models like BloombergGPT, FinGPT prioritizes accessibility and adaptability, allowing users to fine-tune the model with new data for timely market insights.

Key Features:

• Open-Source: Provides accessible alternative to proprietary financial LLMs.
• Adaptable: Enables swift fine-tuning to incorporate new data, costing less than $300 per fine-tuning.
• RLHF (Reinforcement Learning from Human Feedback): Incorporates RLHF to learn individual preferences, similar to ChatGPT and GPT-4.

FinGPT-Forecaster

FinGPT offers a Financial Sentiment Analysis tool. This tool allows users to receive a well-rounded analysis of a company and a prediction for next week's stock price movement.

To use the FinGPT-Forecaster, the user must provide:

• Ticker symbol (e.g. AAPL, MSFT, NVDA)
• The day from which you want the prediction to happen (yyyy-mm-dd)
• The number of past weeks where market news are retrieved
• Whether to add the latest basic financials as additional information

FinGPT V3 for Financial Sentiment Analysis

FinGPT V3 series are LLMs finetuned with the LoRA method on the News and Tweets sentiment analysis dataset which achieve the best scores on most of the financial sentiment analysis datasets with low cost.

FinGPT v3.3 use llama2-13b as base model; FinGPT v3.2 uses llama2-7b as base model; FinGPT v3.1 uses chatglm2-6B as base model.

How does FinGPT work?

FinGPT operates on a full-stack framework with five layers:

1. Data Source Layer: Ensures comprehensive market coverage through real-time information capture.
2. Data Engineering Layer: Processes real-time NLP data, tackling the challenges of temporal sensitivity and low signal-to-noise ratio.
3. LLMs Layer: Employs fine-tuning methodologies like LoRA to adapt to the dynamic nature of financial data.
4. Task Layer: Executes fundamental tasks that serve as benchmarks for performance evaluations.
5. Application Layer: Showcases practical applications and demos of FinGPT in the financial sector.

Open-Source Base Models used in FinGPT:

• Llama-2
• Falcon
• MPT
• Bloom
• ChatGLM2
• Qwen
• InternLM

Who is FinGPT for?

FinGPT is designed for:

• Financial analysts and researchers: Seeking tools for sentiment analysis, financial forecasting, and risk analysis.
• Developers: Building AI-powered financial applications and robo-advisors.
• Institutions: Aiming to leverage open-source LLMs for financial data analysis and market monitoring.

Why choose FinGPT?

• Democratization of Financial Data: Offers an accessible alternative to closed-source financial LLMs.
• Adaptability and Speed: Enables swift fine-tuning with new data for timely market insights.
• Community-Driven: Benefits from contributions and improvements from the open-source community.

FinGPT Ecosystem

• FinGPT-RAG: Enhances financial sentiment analysis by optimizing information depth and context through external knowledge retrieval.
• FinGPT-FinNLP: Provides a playground for all people interested in LLMs and NLP in Finance.
• FinGPT-Benchmark: Introduces a novel Instruction Tuning paradigm optimized for open-source Large Language Models (LLMs) in finance

Tutorials:

• [Training] Beginner’s Guide to FinGPT: Training with LoRA and ChatGLM2–6B One Notebook, $10 GPU
• Understanding FinGPT: An Educational Blog Series
• FinGPT: Powering the Future of Finance with 20 Cutting-Edge Applications
• FinGPT I: Why We Built the First Open-Source Large Language Model for Finance
• FinGPT II: Cracking the Financial Sentiment Analysis Task Using Instruction Tuning of General-Purpose Large Language Models