Гиперпараметры — конфигурационные значения, задаваемые до или во время обучения и управляющие тем, как алгоритм машинного обучения учится. Примеры включают learning rate, число деревьев, batch size, силу регуляризации или выбор архитектуры модели. В отличие от параметров модели, таких как веса, гиперпараметры не выучиваются напрямую из обучающих примеров.
A parameter of a machine learning algorithm. Examples include the number of trees to learn in a decision forest or the step size in a gradient descent algorithm. Values of Hyperparameters are set before training the model and govern the process of finding the parameters of the prediction function, for example, the comparison points in a decision tree or the weights in a linear regression model. For more information, see the Hyperparameter article on Wikipedia.
The variables that you or a hyperparameter tuning service adjust during successive runs of training a model. For example, learning rate is a hyperparameter. You could set the learning rate to 0.01 before one training session. If you determine that 0.01 is too high, you could perhaps set the learning rate to 0.003 for the next training session. In contrast, parameters are the various weights and bias that the model learns during training. See Linear regression: Hyperparameters in Machine Learning Crash Course for more information.