smile.hpo
Hyperparameter optimization.
Attributes
Members list
Type members
Classlikes
Implicit operations on Hyperparameters.
Implicit operations on Hyperparameters.
Attributes
- Supertypes
-
class AnyValtrait Matchableclass Any
Value members
Concrete methods
Executes Bayesian hyperparameter optimization.
Executes Bayesian hyperparameter optimization.
Value parameters
- acquisition
-
the acquisition function strategy.
- exploration
-
exploration parameter (ξ for EI/PI, κ for UCB).
- hp
-
the hyperparameter configuration space.
- initialTrials
-
the number of initial random trials (-1 for heuristic default).
- kernel
-
the Mercer covariance kernel for GP surrogate modeling.
- maxTrials
-
the total number of evaluations to perform.
- maximize
-
true to maximize the objective, false to minimize.
- objective
-
the objective function evaluating a configuration.
Attributes
- Returns
-
the optimization result containing best parameters, value, and trial history.
Creates a new hyperparameter search space.
Creates a new hyperparameter search space.
Attributes
- Returns
-
the hyperparameter search space builder.
Implicits
Implicits
Implicit operations on Hyperparameters.
Implicit operations on Hyperparameters.