Documentation

Welcome to FinWorld - Your comprehensive guide to financial AI development

Welcome to FinWorld

FinWorld is a comprehensive framework for financial AI research and development. It provides a unified platform for data processing, model training, backtesting, and deployment of financial AI systems.

Our framework supports multiple AI paradigms including reinforcement learning, machine learning, and rule-based approaches, making it suitable for various financial applications from trading to portfolio management.

Architecture Guide

Learn about FinWorld's seven-layer architecture and how components interact to build robust financial AI systems.

Learn More

Financial Tasks

Comprehensive overview of the four core financial AI tasks: forecasting, trading, portfolio management, and LLM applications.

Learn More

Configuration Guide

Understand how to configure FinWorld using YAML files and the mmengine configuration system.

Learn More

API Reference

Complete API documentation for all FinWorld components, classes, and methods.

Learn More

Tutorials

Step-by-step tutorials to get you started with FinWorld development and deployment.

Learn More

Examples

Practical examples and code samples to help you implement FinWorld in your projects.

Learn More

Installation

FinWorld requires Python 3.11+ and can be installed using conda and make commands. Here's the complete installation process:

Requirements

  • Python 3.11+
  • CUDA 12.4+ (for GPU acceleration)
  • Conda or Miniconda
  • 16+ GB RAM (recommended for large datasets)

1. Create Conda Environment

conda create -n finworld python=3.11
conda activate finworld

2. Install Dependencies

# Install base dependencies
make install-base

# Install browser automation tools
make install-browser

# Install VERL framework
make install-verl

Alternative Installation with Poetry

# Install Poetry
pip install poetry

# Install dependencies
poetry install

Quick Start

Get started with FinWorld in just a few steps:

1. Download Financial Data

# Download DJ30 data (example)
python scripts/download/download.py --config configs/download/dj30/dj30_fmp_price_1day.py
python scripts/download/download.py --config configs/download/dj30/dj30_fmp_price_1min.py

2. Train RL Trading Models

# Train PPO trading models for multiple stocks
CUDA_VISIBLE_DEVICES=0 python scripts/rl_trading/train.py --config=configs/rl_trading/ppo/AAPL_ppo_trading.py

3. Train Portfolio Models

# Train PPO portfolio models for different indices
CUDA_VISIBLE_DEVICES=0 python scripts/rl_portfolio/train.py --config=configs/rl_portfolio/ppo/dj30_ppo_portfolio.py

4. Use Pre-built Scripts

# Run example scripts
bash examples/ppo_trading.sh
bash examples/ppo_portfolio.sh
bash examples/download.sh