Optuna Backend#
Hyperactive provides wrappers for Optuna’s powerful samplers, giving you access to Optuna’s algorithms with Hyperactive’s unified interface.
Available Optimizers#
from hyperactive.opt.optuna import (
TPEOptimizer, # Tree-Parzen Estimators
CmaEsOptimizer, # CMA-ES evolution strategy
GPOptimizer, # Gaussian Process
NSGAIIOptimizer, # Multi-objective (NSGA-II)
NSGAIIIOptimizer, # Multi-objective (NSGA-III)
QMCOptimizer, # Quasi-Monte Carlo
RandomOptimizer, # Random sampling
GridOptimizer, # Grid search
)
Example: TPE with Optuna#
Example with Optuna TPE:
from hyperactive.opt.optuna import TPEOptimizer
optimizer = TPEOptimizer(
search_space=search_space,
n_iter=50,
experiment=objective,
)
When to Use Optuna Backend#
The Optuna backend is useful when:
You want access to Optuna’s well-tested sampler implementations
You’re familiar with Optuna and want similar behavior
You need specific Optuna features like CMA-ES or NSGA-II/III
For most use cases, the native GFO optimizers provide equivalent functionality with the same interface.