Team: Melody Llinas, Bruke Amare, Udom An, Miya Reese, & Mark Zhang
This project is for the 2024 UNH ML-Hackathon.
This repository contains the files for the "Navy Ship" chatbot, an interactive Streamlit-based application designed to answer questions based on a collection of Navy-related documents. It uses Llama 3.2 for semantic search, language processing, and spell correction to provide relevant responses.
Ensure you have the required libraries installed by running:
pip install streamlit torch transformers pymilvus spellchecker requests streamlit-shadcn-ui PyPDF2 langchain sentence-transformers
or
pip install -r requirements.txt
Note: You MUST do this before running the streamlit app.
Run the following commands in your terminal:
brew install ollama
ollama serve
ollama pull llama3.2
Now leave this terminal running, and open another terminal.
Start Streamlit App: To launch the chatbot, run the following command:
streamlit run Pro_Code_Tier.py
Interact with the app here: https://naval-chatbot.streamlit.app/