Also known as: Pre-Trained Models Β· Pre trained Model
A pre-trained model is an AI model trained on a broad or large dataset before being adapted for a specific task. It can provide reusable general capabilities that are later refined through fine-tuning, prompting, or other customization methods.
Although this term could refer to any trained model or trained embedding vector, pre-trained model now typically refers to a trained large language model or other form of trained generative AI model. See also base model and foundation model.
βa component of the training stage in which a model learns general patterns, features, and relationships from vast amounts of unlabeled data, such as through self-supervised learning. Pre-training can equip models with knowledge of general features or patterns which may be useful in downstream tasks, and can be followed with additional training or fine-tuning that specializes the model for a specific downstream task.β