Best GitHub Repositories to Learn LLM & Generative AI

Build a GPT-style LLM from scratch

The official code for building a large language model from scratch in PyTorch, step by step.

Best for: Developers who want to understand LLMs at the code level.

A roadmap and notebooks for LLMs

A course with roadmaps and Colab notebooks covering LLM fundamentals, fine-tuning, and deployment.

Best for: Engineers moving into applied LLM work.

The definitive prompt engineering guide

A comprehensive, frequently updated guide to prompt engineering techniques and research.

Best for: Anyone building with LLMs who wants better outputs.

A 21-lesson GenAI course by Microsoft

A free, structured course teaching the fundamentals of building generative-AI applications.

Best for: Beginners entering generative AI.

A curated list of LLM resources

A curated collection of papers, models, tools, and resources about large language models.

Best for: Learners tracking the fast-moving LLM space.

Run LLMs locally with one command

A tool to download and run open LLMs (Llama, Mistral, etc.) locally on your machine.

Best for: Developers experimenting with open models privately.

State-of-the-art models, one API

Hugging Face Transformers — the library powering most modern NLP and LLM work.

Best for: Developers building with modern AI models.

Data framework for LLM applications

LlamaIndex — a framework for connecting custom data to LLMs (RAG).

Best for: Developers building RAG systems.

A self-hosted UI for LLMs

A feature-rich, self-hostable web interface for running and chatting with LLMs.

Best for: Devs wanting a private LLM interface.

Curated datasets for LLM training

A curated collection of datasets and resources for training and fine-tuning LLMs.

Best for: Practitioners fine-tuning models.

Build LLM apps in JavaScript/TS

The JavaScript/TypeScript version of LangChain for building LLM-powered apps.

Best for: JS/TS devs building AI apps.

High-throughput LLM serving

vLLM — a fast, memory-efficient engine for serving large language models.

Best for: Engineers deploying LLMs in production.

Run local LLMs on your hardware

GPT4All — an ecosystem to run open LLMs locally on consumer hardware.

Best for: Users wanting private local LLMs.