fine-tuning
Extra training that adapts an existing model to one specific job, using your own examples.
Training a model from scratch is very expensive. Fine-tuning starts from a model that already works and trains it a little more on a smaller set of examples from your own work. The model keeps its general skill and gets better at your specific task, style or labels. It needs far less data and computing power.
Think of it as
Like hiring an experienced accountant and spending a week showing them how your company files expenses. You are not teaching accounting, just your way of doing it.
Example
A support team has thousands of past tickets, each already tagged with the right team. Fine-tuning a general model on them teaches it that team's own categories, so it tags new tickets the way that team would.