Class HyperbandPruner

Namespace

OptunaSharp.Pruners

Assembly

OptunaSharp.dll

Pruner using Hyperband. Port of optuna/pruners/_hyperband.py: HyperbandPruner.

As SuccessiveHalving (SHA) requires the number of configurations n as its hyperparameter, for a given finite budget B all the configurations have the resources of B/n on average. Hyperband attacks this trade-off by trying different n values for a fixed budget. Hyperband runs multiple SuccessiveHalvingPruners ("brackets"); the number of brackets is floor(log_eta(maxResource/minResource)) + 1.

HyperbandPruner computes the bracket ID for each trial from study_name and the trial number (CRC-32), so specify studyName to make the pruning algorithm reproducible.

public class HyperbandPruner : BasePruner

Inheritance

object

BasePruner

HyperbandPruner

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Constructors

HyperbandPruner(long, long?, int, int)

Configures resource brackets whose successive-halving competitions have separate trial histories.

public HyperbandPruner(long minResource = 1, long? maxResource = null, int reductionFactor = 3, int bootstrapCount = 0)

Parameters

minResource long

The minimum resource allocated to a trial (r in the paper). See SuccessiveHalvingPruner.

maxResource long?

The maximum resource allocated to a trial, or null for Optuna's "auto": estimated from the first completed trial's largest reported step (+1); no trials are pruned until determined. This value should match the maximum iteration steps.

reductionFactor int

Reduction factor of promotable trials (eta in the paper).

bootstrapCount int

Number of trials required in a rung before any trial can be promoted. Incompatible with automatic maxResource.

Exceptions

ValueError

Positive bootstrap is combined with automatic maximum resource.

Methods

Prune(Study, FrozenTrial)

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

Note that this method is not supposed to be called by library users. Instead, Report(double, long) and ShouldPrune() provide user interfaces to implement pruning mechanism in an objective function.

public override bool Prune(Study study, FrozenTrial trial)

Parameters

study Study

Study object of the target study.

trial FrozenTrial

FrozenTrial object of the target trial. Take a copy before modifying this object.

Returns

bool

A boolean value representing whether the trial should be pruned.

Exceptions

RuntimeError

The study has multiple objectives and the pruning rule requests its scalar direction.

ValueError

Deferred bracket initialization rejects invalid resource, reduction, or bootstrap settings.

InvalidOperationException

Initialized bracket allocation budgets are inconsistent.

See Also

Official Optuna HyperbandPruner reference.