Implemented optimizers#

The following implementations of optimizers can found under this extension package:

Nevergrad optimizers#

These derivative-free global optimizers are ported from the nevergrad Python library:

In order to use this optimizers you need to install the nevergrad dependency:

pip install nevergrad

Name	Keyword argument	Function	Reference
Differential Evolution	`"de_nevergrad"`	`de_nevergrad()`	nevergrad.optimizers.DifferentialEvolution
PSO	`"pso_nevergrad"`	`pso_nevergrad()`	nevergrad.optimizers.ConfPSO

These derivative-free global optimizers are ported from the evosax Python library:

In order to use this optimizers you need to install the evosax dependency:

pip install evosax

Note

The evosax library is only available for Linux and Unix systems.

Name	Keyword argument	Function	Reference
CMAES	`"cmaes"`	`cmaes()`	evosax.strategies.cma_es
PSO	`"pso"`	`pso()`	evosax.strategies.pso
Simulated Annealing	`"simanneal"`	`simanneal()`	evosax.strategies.sim_anneal
Differential Evolution	`"de"`	`de()`	evosax.strategies.de

These derivative-free global optimizers are ported from the optuna Python library:

In order to use this optimizers you need to install the optuna dependency:

pip install optuna

Name	Keyword argument	Function	Reference
Tree-structured Parzen Estimator	`"tpe_sampler"`	`tpe_sampler()`	optuna.samplers.TPESampler

These derivative-free global optimizers are ported from the optax Python library:

In order to use this optimizers you need to install the optax dependency:

pip install optax

Name	Keyword argument	Function	Reference
Adam	`"adam"`	`adam()`	optax.adam
SGD	`"sgd"`	`sgd()`	optax.sgd