Glossary

Hyperparameter — Glossary

A configuration choice set before learning parameters, such as learning rate or number of layers.

Reviewed:

Full definition

Hyperparameters are tuned via validation performance, search methods, or heuristics. Decoding settings like temperature are inference hyperparameters for generative models.

Related encyclopedia topic

Hyperparameters — Hyperparameters are training and model choices set before learning parameters—learning rate, batch size, depth, and regularization strength among them.

← Back to glossary · Learning hub