LocScalePTM.optim()

LocScalePTM.optim()#

LocScalePTM.optim(exclude_roles=None, exclude_params=None, swap_pairs=None, stopper=None, progress_bar=False, test_for_positive_definiteness=False, update_parameters_inplace=False, **kwargs)[source]#

Optimize model parameters using the selected optimizer.

Parameters:

  • exclude_roles (list[Union[Literal['hyperparam', 'transformation_coef'], str]] | None) – Roles to exclude from optimization. (default: None)

  • exclude_params (Sequence[str] | None) – Specific parameter names to exclude. (default: None)

  • swap_pairs (Sequence[SwapSpec] | None) – Swap specifications used to temporarily replace distribution parts during optimization. (default: None)

  • stopper (Stopper | None) – Optional stopper controlling optimization iterations. (default: None)

  • progress_bar (bool) – Whether to show a progress bar. (default: False)

  • test_for_positive_definiteness (bool) – If True, test Fisher information matrices for positive definiteness. (default: False)

  • update_parameters_inplace (bool) – If True, apply optimized parameters to the model. Otherwise restore the previous state after optimization. (default: False)

  • **kwargs – Forwarded to the underlying optimizer.

Return type:

OptimResult

Returns:

OptimResult – Result of the optimization run.

Contents