notebooks

Why we included this project

Fine-tuning an open model is rarely a one-shot affair. This repository collects working notebooks that take you through the whole pipeline, from formatting a dataset to training and running inference, across text, vision, audio, embedding, and TTS models. With more than 250 notebooks organized by model, you can usually find one that matches the architecture you have in mind, whether that's a small chat model on a free Colab GPU or a GRPO-style RL run on a math benchmark. The real payoff is executable code instead of a blank template: the notebooks cover data preparation, supervised fine-tuning, inference, and reinforcement learning, which cuts down the trial and error of adapting an open model to your own data. Teams wanting a repeatable starting point for parameter-efficient fine-tuning will find these a practical reference to build from.