datawhalechina/leegenai-tutorial

What it solves

This project provides a structured, accessible entry point for beginners to understand Generative Artificial Intelligence. It demystifies complex AI concepts by organizing high-quality lecture materials into a clear, step-by-step learning path.

How it works

The tutorial aggregates and organizes lecture videos, slides, and course materials from Professor Hung-yi Lee’s courses at National Taiwan University (specifically the 2024 Spring "Introduction to Generative AI" and 2025 Spring "Machine Learning" courses). It breaks down the curriculum into logical modules covering everything from basic definitions to advanced topics like Transformer architecture, AI agents, and model safety.

Who it’s for

It is designed for students, developers, and AI enthusiasts who are new to the field of generative AI and prefer a clear, academic, yet easy-to-understand teaching style.

Highlights

• Comprehensive coverage of Generative AI, including prompt engineering, LLM training stages, and AI agents.
• Includes deep dives into technical foundations like Transformer architecture and model interpretability.
• Addresses practical and ethical concerns such as model evaluation, safety, and adversarial attacks.
• Provides insights into cutting-edge topics like image generation (Sora-related principles) and advanced model techniques like Speculative Decoding.
• Offers downloadable PDF versions of course materials for offline study.