Training an existing foundation model further on a smaller, specialized dataset so it performs better on a particular task or domain.
Fine-Tuning is a machine learning technique in which an existing, general-purpose foundation model goes through an additional round of training on a specialized dataset.
An organization can take an open-source Large Language Model and fine-tune it on its own medical records, legal contracts, or engineering schematics, at far lower cost than training a model from scratch.
The result is a specialized model that works with the organization's own domain language and business logic, and it becomes a central asset in a Private AI deployment.
See also Machine Learning, Large Language Model, Private AI, Model Inference and MLOps.