Fine-Tuning
Fine-tuning is a second, lighter round of training applied to a model that has already been trained on a large, general collection of data. Instead of building a model from scratch, you start from an existing one and continue training it on a smaller, more focused set of examples so it adapts to a particular task, domain, or language. Because the model already carries a lot of general capability, fine-tuning usually needs far less data and computing power than training from the beginning, which is part of why it is attractive for low-resource languages where little digital text exists.