Overview

This personal project brings together cutting-edge technology to create a local assistant chatbot that’s both powerful and secure. Using freely available large language models (LLMs), this chatbot runs on your own machine, ensuring your data stays private. It’s specifically designed to help with construction materials, tapping into a detailed database you provide to deliver precise information about various material properties.

Key Features

• Local LLM Execution: The chatbot runs the Llama.cpp model locally, keeping your data safe and performance smooth.
• Custom Database Integration: Connects to your personal database with detailed information on construction materials, covering everything from physical and thermal properties to mechanical characteristics.
• Streamlit Interface: Offers an easy-to-use, conversational interface, making interactions with the chatbot intuitive and user-friendly.

Technical Highlights

• Python Development: Built entirely in Python
• Machine Learning Model: Uses the Mistral model from Hugging Face
• Database Management: Incorporates a detailed database for accurate information retrieval and can provide fallback answers from the web if the data isn’t available.

Why It Matters

This project is a practical tool designed to manage construction materials efficiently. By combining a local LLM with a customized database, it ensures your data remains private and secure. It allows easy access to detailed information about materials’ properties, which is crucial for making informed decisions on construction sites. Plus, the chatbot’s ability to offer alternative answers from the web when needed adds flexibility, making it a versatile tool for various applications.