Optuna

Class resolvers for Optuna.

pruner_resolver = <class_resolver.api.ClassResolver object>

A resolver for optuna.pruners.BasePruner subclasses.

Building on the simple example from the Optuna website’s homepage, you can parametrize optuna.create_study() with a pruner instantiated with class_resolver.

import optuna
from class_resolver import Hint
from class_resolver.contrib.optuna import pruner_resolver
from optuna.pruner import BasePruner

def objective(trial):
    x = trial.suggest_float('x', -10, 10)
    return (x - 2) ** 2

def optimize_study(pruner: Hint[BasePruner] = None):
    study = optuna.create_study(
        pruner=pruner_resolver.make(pruner),
    )
    study.optimize(objective, n_trials=100)
    return study

study = optimize_study(pruner="median")
study.best_params  # E.g. {'x': 2.002108042}

sampler_resolver = <class_resolver.api.ClassResolver object>

A resolver for optuna.samplers.BaseSampler subclasses.

Building on the simple example from the Optuna website’s homepage, you can parametrize optuna.create_study() with a sampler instantiated with class_resolver.

import optuna
from class_resolver import Hint
from class_resolver.contrib.optuna import sampler_resolver
from optuna.sampler import BaseSampler

def objective(trial):
    x = trial.suggest_float('x', -10, 10)
    return (x - 2) ** 2

def optimize_study(sampler: Hint[BaseSampler] = None):
    study = optuna.create_study(
        sampler=sampler_resolver.make(sampler),
    )
    study.optimize(objective, n_trials=100)
    return study

study = optimize_study(sampler="TPE")
study.best_params  # E.g. {'x': 2.002108042}