UNF Tour Guide Chatbot

A conversational AI chatbot that provides information about the University of North Florida campus, locations, buildings, and facilities.

Prerequisites

• Python 3.x installed
• Ollama installed (Download here)

Models

• Testing: Mistral (lightweight, fast iteration)
• Deployment: Llama2 (more capable, production-ready)

Setup (First Time Only)

1. Install Python packages:

   pip install -r requirements.txt
   

   Or manually install: pip install ollama pydantic torch

---

Running the Chatbot - Main Method

(Coming soon - model training in progress)

---

Running the Chatbot - Testing Method

Prerequisites for Testing

• The Mistral model downloaded:

  ollama pull mistral
  

  This downloads the AI model (~4GB). Only do this once.

Step 1: Start Ollama Service

Open a PowerShell terminal and run:

ollama serve

Leave this running in the background. You should see "Listening on localhost:11434"

Step 2: Run the Testing Chatbot

Open a new PowerShell terminal and run:

& "tour_bot\Scripts\Activate.ps1"
python testing-chat-bot.py

The chatbot will start and display:

============================================================
UNF Tour Guide Chatbot
============================================================
Type 'quit' or 'exit' to end the conversation
Type 'building <number>' to change buildings

Using the Testing Chatbot

• Type your questions about UNF campus
• Examples:
  • "Where is the library?"
  • "Tell me about the student center"
  • "How do I get to the engineering building?"
• The bot will respond with helpful information and maintain conversation history
• Type building <number> to switch buildings (this resets conversation history)
• Type quit, exit, or q to end the conversation

Features

• Conversation Memory: The bot maintains chat history within each building session for more natural conversations
• Building Context: Each building has its own information file for specific details
• CPU/GPU Support: Automatically detects and uses GPU if available (configurable with FORCE_CPU flag)

Troubleshooting

"Error: Please make sure Ollama is running"

• Make sure you have ollama serve running in another terminal
• Check that it's listening on localhost:11434

"Model not found" error

• Run ollama pull mistral to download the model

Slow responses

• First query with Mistral takes longer to load
• Subsequent queries will be faster
• This is normal behavior

GPU Not Being Used

• Check with nvidia-smi to verify your NVIDIA drivers are installed
• Install PyTorch with CUDA support: pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
• Set FORCE_CPU = False in testing-chat-bot.py

Notes

• Ollama must be running continuously while you use the chatbot
• You only need to pull the model once
• Running both Ollama and the chatbot in separate terminals keeps them responsive
• The testing chatbot is set up specifically for UNF campus information