Open Machine Learning (OLML)

A collection of tools and examples for working with AI models and Language Learning Models (LLMs) focused on health-related applications.

Overview

This repository contains various examples and implementations of AI-powered applications with a focus on:

• Retrieval Augmented Generation (RAG) systems
• Medical question answering
• Women's health information processing
• Integration with various LLM providers (OpenAI, etc.)

Key Components

Notebooks

• gen-qa-openai.ipynb: Retrieval Enhanced Generative Question Answering with OpenAI to resolve hallucinations in LLMs
• rag_pubmed.ipynb: Medical Question Answering system using LangChain and Mistral 7B with PubMed data

Python Scripts

• test_langchain.py: Demonstrates LangChain integration with PubMed for medical information retrieval
• womens_health.py: Uses AIConfig to process women's health-related information
• script.py: Main script for running AI-based operations using various LLM providers

Configuration Files

• health.aiconfig.yaml: Configuration for AI models focused on health-related prompts and contexts
• .env: Environment configuration for API keys from different LLM providers

Getting Started

Prerequisites

• Python 3.8+
• API keys for LLM providers:
  • OpenAI
  • Google (optional)
  • Hugging Face (optional)

Installation

1. Clone this repository:

git clone https://github.com/yourusername/OLML.git
cd OLML

2. Create and activate a virtual environment (recommended):

python -m venv venv
source venv/bin/activate  # On Windows, use: venv\Scripts\activate

3. Install the required dependencies:

pip install -r requirements.txt  # Note: You may need to create this file

4. Configure your API keys:
   • Rename .env.example to .env (if available)
   • Add your API keys to the .env file

Usage

Running Notebooks

Open the Jupyter notebooks in your preferred environment:

jupyter notebook

Then navigate to either gen-qa-openai.ipynb or rag_pubmed.ipynb.

Using the Python Scripts

Run the test_langchain.py script:

python test_langchain.py

Run the women's health script:

python womens_health.py

License

[Include license information here]

Acknowledgments

• This project uses various AI and ML libraries including LangChain, OpenAI, and others
• Medical data sources include PubMed and women's health resources