In deep learning, fine-tuning is the process of adapting a computational model trained for one task to perform a different, usually more specific, task. It is considered a form of transfer learning, as it reuses knowledge learned from the original training objective.