Pre-training

Pre-training is a machine learning strategy in which a model is first trained on a sizable dataset to learn general features and patterns. The pre-trained model gets further tuned on specific tasks with smaller datasets, leveraging the learned information for improved task-specific performance.