Health Insurance Document AI is a generative AI-powered tool designed to read and interpret health insurance PDFs. It provides real-time answers on coverage, pricing, and other details directly from the document. This project utilizes advanced machine learning models and technologies to offer a seamless user experience.
- AI Model: Utilizes Llama from Hugging Face, fine-tuned with Unsloth AI on a comprehensive health insurance plans corpus.
- RAG Pipeline: Implements a retrieval-augmented generation (RAG) pipeline with Chroma DB for accurate and relevant information retrieval.
- User Interface: Developed with Flask for an intuitive and user-friendly interaction.
- Local Hosting: Hosted locally on Ollama to ensure privacy and high performance.
Here are some resources to help you understand and get started with Generative AI:
- Hugging Face: Comprehensive guide to models and tools. 🧑💻📚
- Unsloth AI: Detailed documentation for fine-tuning and usage. 📖🔧
- IBM LLM Course: In-depth video series on large language models. 🎓🎥
- LangChain Course: Taught by Andrew Ng and LangChain creator Harry Chase. 🌟📚
- CodeBasics: Practical tutorials on GenAI. 💻🎓
- Generative AI with LLMs: Essential course for understanding GenAI fundamentals. 🚀📖
Contributions are welcome! If you have suggestions or improvements, please fork the repository and submit a pull request. For major changes, please open an issue to discuss what you would like to change.
For any questions or feedback, feel free to reach out to me at [email protected]
Special thanks to:
- Harry Chase for his work on LangChain
- Unsloth AI for their fine-tuning tools
- Ollama for local model hosting
- opm.gov for valuable insights and support