Hyperparameters

In machine learning, a hyperparameter is a parameter whose value is used to control the learning process. By contrast, the values of other parameters (typically node weights) are derived via training.

The time required to train and test a model can depend upon the choice of its hyperparameters.

They are important to performance of a model, but tuning them can be ignored if the amount of computing power available to the researcher is not enough to do multiple rounds of training since it is mostly trial-and-error.

A good starting point would be to use hyperparameters used in similar tasks, such as a language model for a related language.