Technical
Trained model
The model artifact produced by training on a dataset.
Definition
The resulting trained machine-learning artifact produced from applying a training algorithm to a training dataset, encompassing model parameters, weights, architecture, and potentially associated metadata (e.g., hyperparameters, versioning). The Guidelines address allocation of rights, access, and obligations related to trained models in contracts.
Related Terms
Training dataset
Dataset used to train a machine learning model....
Model artefacts
Learned model components such as weights or pre-trained models....
Model
A trained algorithmic artifact that performs tasks using learned parameters....
Model Training
Process of improving AI systems using data....
Models
Trained algorithms used to generate inferences or predictions....