Exercises and Labs for EECE.4860/5860 Introduction to Generative AI at UMass Lowell
Exercise: a standalone, complete implementation that introduces a concept or method, and provides an executable example to demonstrate the concept or method. No code addition or modification is needed.
Lab: a standalone but incomplete Jupyter notebook that focuses on a concept discussed in the class. Students are expected to complete the notebook by adding code or modifying parts of the existing code, execute the notebook and analyze the results. A lab report is required as part of the lab submission. Lab grades will be based on the code submission and lab reports.
- GenAI Lab Setup Guide
- Handwritten Digits Recognition
- LLM basics with Hugging Face
- Gemma model fine tuning
- Simple LLM inference
- Toxicity Detector by Meta
- Lab 1: Use of LLM, Prompt Engineering and Model Finetuning
- Lab 2: LoRA Fine-Tuning, Model Evaluation and RLHF
- Lab 3: RAG and Langchain AI agent
A collection of FAQs are available.