Single Search Method

The single search method does a very minimal “search”: it trains a single hyperparameter configuration for the number of units specified by max_length (see Training Units) and then performs validation. This method is useful for testing or for training a single model configuration until convergence.

See Experiment Configuration.