FineTuningLLMs

Why we included this project

This is Daniel Godoy's official companion to his book on fine-tuning LLMs, but the notebooks stand on their own as a tutorial. Each chapter is a Jupyter notebook you can open directly in Google Colab and run on a free GPU, walking through loading a quantized model, configuring LoRA adapters, formatting your dataset, training with SFTTrainer, and exporting a GGUF file you can serve locally with Ollama or llama.cpp. If you already work with PyTorch and the Hugging Face stack and want to actually fine-tune a model on a single consumer GPU rather than only read theory, the code here is a concrete route through the full process. The notebooks also make a handy reference to adapt when you start your own fine-tuning work.