LLM Course

LLM Course is a free, open-source learning resource by Maxime Labonne, hosted on GitHub under the Apache License 2.0. It combines structured roadmaps with curated references and Google Colab notebooks covering LLM fundamentals, model building, and application engineering.

What It Is

The course is divided into three parts: LLM Fundamentals (optional math, Python, neural networks, and NLP), The LLM Scientist (building LLMs with current techniques), and The LLM Engineer (creating and deploying LLM-based applications). Each topic pairs a short explanation of key concepts with a list of reference articles, videos, and libraries. The author states the course will always stay free and that it informed his book, the LLM Engineer's Handbook.

What the Curriculum Covers

The Scientist track covers LLM architecture, pre-training, post-training datasets, supervised fine-tuning, preference alignment (DPO, GRPO, PPO), evaluation, quantization, and newer trends such as model merging, multimodal models, interpretability, and test-time compute. The Engineer track covers running LLMs, building vector storage, retrieval augmented generation, advanced RAG, agents, inference optimization, deployment, and security.

Notebooks and Tools

A notebook section lists Colab notebooks and companion articles on fine-tuning (Unsloth, ORPO, DPO, QLoRA, Axolotl), quantization (GGUF, GPTQ, EXL2), and model merging with MergeKit. It also lists one-click tools such as LLM AutoEval, LazyMergekit, LazyAxolotl, and AutoQuant. The author's blog hosts the related technical articles.