Senior Project Notice
This repository was created for a senior project in ENGT 375 Applied Machine Learning at Old Dominion University. It is provided for educational and research demonstration purposes only. It is not intended for production use, security filtering, or making real-world spam/phishing decisions. Always use established security tools for operational email protection.
Documentation
Beginner-friendly guides for the Liquid AI spam classifier project.
Guides
| # | Guide | What You'll Learn |
|---|---|---|
| 1 | What is Liquid AI? | The Liquid AI model family and how it differs from traditional LLMs |
| 2 | What is LoRA? | How LoRA makes fine-tuning affordable on a laptop |
| 3 | Training Guide | Step-by-step walkthrough of fine_tune.py |
| 4 | HuggingFace TRL + PEFT Reference | Key APIs and tools used in this project |
| 5 | Deployment Guide | Running the app on your Mac |
| 6 | Setup Guide | Environment setup from scratch |
| 7 | Code Sources & References | Every source, citation, and finding for paper writing |
| 8 | GGUF Conversion Guide | How to convert your LoRA adapter to GGUF for Ollama, LM Studio, llama.cpp |
Quick Commands
| Action | How |
|---|---|
| Retrain | Double-click retrain.command (consolidated; replaces the older fast/full split) |
| Evaluate only | Double-click evaluate.command |
| Launch app | Double-click launch UI.command |
| Open notebook | Double-click launch-notebook.command |
| Build GGUF | Double-click BuildGGUF.command (see Guide 8) |
See Training Guide for details on the new 3-class (spam/ham/phishing) datasets.
Reading Order
If you are new to this project, read them in order (1 → 6). Guide 7 is a reference for paper writing — use it when citing sources.