hnn-llm is a tool developed to enhance the accessibility of information related to the Human Neocortical Neurosolver (HNN). It aims to address the challenge of efficiently finding relevant information within extensive HNN documentation and related resources.

Purpose

As someone deeply interested in computational neuroscience and HNN, I often struggled to navigate extensive documentation to find precise information during my internship as a Google Summer of Code 2024 contributor at INCF under the HNN-Core project. To address this challenge, I developed this project to streamline the search process using state-of-the-art technologies.

Important

This project was created independently and is not part of my Google Summer of Code Internship.

Features

• Retrieval-Augmented Generation (RAG) Model: Utilizes the GroqAPI and Llama3 8b model to provide accurate responses based on user queries.
• Vector Search and Embedding: Employs Chroma DB for effective vector search and document embedding.
• User Interface: Built with Streamlit to offer an intuitive and interactive user experience.
• Data Sources: Retrieves and processes information from the official HNN website and the HNN-Core documentation site.

License

This project is licensed under the BSD 3-Clause License. See the LICENSE file for more details.