optuna

See also

Full documentation with examples can be found here: documentation page

Optuna only functionality.

class hpoflow.optuna.SignificanceRepeatedTrainingPruner(alpha=0.1, n_warmup_steps=4)[source]

Bases: BasePruner

Pruner which uses statistical significance as an heuristic for decision-making.

Pruner to use statistical significance to prune repeated trainings like in a cross validation. As the test method a t-test is used. Our experiments have shown that an aplha value between 0.3 and 0.4 is reasonable.

Constructor.

Parameters:

  • alpha (float) – The alpha level for the statistical significance test. The larger this value is, the more aggressively this pruner works. The smaller this value is, the stronger the statistical difference between the two distributions must be for Optuna to prune. alpha must be 0 < alpha < 1.

  • n_warmup_steps (int) – Pruning is disabled until the trial reaches or exceeds the given number of steps.

prune(study, trial)[source]

Judge whether the trial should be pruned based on the reported values.

Parameters:
Return type:

bool