How to set c1, c2 and w parameters of Nevergrad.PSO algo?

Ray Libraries (Data, Train, Tune, Serve) Ray Tune
1

Can anyone guide me how to initialize Nevergrad.PSO with custom values for c1, c2 and w to control exploration and exploitation? Here are relevant doc strings, which dont talk about this info.

ng.optimizers.ConfiguredPSO(
    transform: str = 'identity',
    wide: bool = False,
    popsize: Union[int, NoneType] = None,
) -> None
Docstring:     
`Particle Swarm Optimization <https://en.wikipedia.org/wiki/Particle_swarm_optimization>`_
is based on a set of particles with their inertia.
Wikipedia provides a beautiful illustration ;) (see link)


Parameters
----------
transform: str
    name of the transform to use to map from PSO optimization space to R-space.
wide: bool
    if True: legacy initialization in [-1,1] box mapped to R
popsize: int
    population size of the particle swarm. Defaults to max(40, num_workers)

and

Init signature:
NevergradSearch(
    optimizer: Union[NoneType, nevergrad.optimization.base.Optimizer, nevergrad.optimization.base.ConfiguredOptimizer] = None,
    space: Union[Dict, nevergrad.parametrization.core.Parameter, NoneType] = None,
    metric: Union[str, NoneType] = None,
    mode: Union[str, NoneType] = None,
    points_to_evaluate: Union[List[Dict], NoneType] = None,
    max_concurrent: Union[int, NoneType] = None,
    **kwargs,
)
2

Found a useful suggestion from Nevergrad engineers at Nevergrad Users : I'm new to Nevergrad

3

Thanks for sharing this!