Licenses

McCoy Health is committed to delivering cutting-edge health and medical education through innovative technologies. By utilizing open-source AI and Large Language Models (LLMs) such as LLaMA 2, BLOOM, GPT-J-6B, and Falcon Series, we can create more robust and effective educational tools. These models, with their powerful text generation and multilingual capabilities, enable us to develop interactive and personalized learning experiences. Specifically, the medical-specific LLMs like Med-PaLM allow us to integrate accurate and up-to-date medical knowledge, ensuring that our users receive the best possible education and support. Utilizing these advanced technologies not only enhances the quality of our offerings but also ensures that we remain at the forefront of healthcare innovation.

LLaMA 3 (Meta)

  • License: Custom License (Free for use under 700M users, outputs cannot be used to train other LLMs)

  • Details: Suitable for a variety of natural language generation tasks, including chatbots and programming tasks​ (Learn R, Python & Data Science Online)​​ (The GitHub Blog)​.

BLOOM (Hugging Face)

  • License: Open Access License

  • Details: An autoregressive LLM trained on a vast multilingual dataset, providing capabilities in text generation, summarization, translation, and more​ (Learn R, Python & Data Science Online)​​ (Unite.AI)​.

GPT-J-6B

  • License: Apache 2.0

  • Details: An open-source reproduction of OpenAI's GPT-3, suitable for various natural language processing tasks​ (GitHub)​.

Falcon Series (TII)

  • License: Apache 2.0

  • Details: Models such as Falcon-40B and Falcon-7B designed for text generation, translation, and question answering​ (GitHub)​​ (Unite.AI)​.

MPT-7B (MosaicML)

  • License: Apache 2.0, CC BY-SA-3.0

  • Details: A GPT-style, decoder-only transformer model optimized for various tasks including chatbots and instruction following​ (Unite.AI)​​ (The GitHub Blog)​.

OpenLLaMA (Berkeley AI Research)

  • License: Apache 2.0

  • Details: An open-source reproduction of Meta’s LLaMA model, trained on one trillion tokens​ (GitHub)​​ (The GitHub Blog)​.

Med-PaLM (Google Research)

  • License: Custom License

  • Details: A medical LLM designed to provide high-quality answers to medical questions, trained on a variety of medical datasets to ensure accuracy and relevance​ (Google AI)​.