happy-llm

Why we included this project

For developers who have used LLMs through APIs or existing tools but never looked under the hood, this free tutorial fills that gap. It starts with NLP basics, then covers the Transformer architecture, pretraining, and fine-tuning, and ends with building a complete LLaMA2-style model from scratch. The lessons pair explanations with code exercises built on mainstream frameworks, so you can actually train a small model instead of just reading about it. Later chapters cover applied techniques like RAG and agents, which makes the jump from model internals to real systems easier. Since it's organized as an open textbook with an online reading site, it works well as a self-paced course or as material for a study group.