Hyperparameter
A hyperparameter is a setting chosen by developers before or during training, such as the learning rate or number of layers, that controls how a model learns rather than being learned itself.
What is Hyperparameter?
Hyperparameters strongly affect results, and finding good values is a major part of model development. They are usually selected by trying different combinations and comparing results on validation data.
How it works
Search methods include grid search, random search and Bayesian optimisation, each training several model variants and keeping the best.
← All Artificial Intelligence & Superintelligence definitions · ← Back to the A–Z definitions