default_integrator module

High level interface for integration

Integrator(f, d, survey_strategy='forward_flat_init', n_iter=10, n_iter_survey=None, n_iter_refine=None, n_points=100000, n_points_survey=None, n_points_refine=None, use_survey=False, device=device(type='cpu'), verbosity=None, trainer_verbosity=None, loss='dkl', flow='pwquad', trainer=None, trainer_options=None, flow_options=None)[source]

High level integration API

This is a factory method that instantiates the relevant Integrator subclass based on the options

Parameters

  • f (function) – the function to integrate

  • d (int) – dimensionality of the integration space

  • survey_strategy (str) – how points are sampled during the survey step: one of 'flat', 'forward', 'adaptive_dkl', 'adaptive_variance', 'forward_flat_init', 'fixed_sample'

  • n_iter (int) – general number of iterations - ignored for survey/refine if n_iter_survey/n_inter_refine is set

  • n_iter_survey (int) – number of iterations for the survey stage

  • n_iter_refine (int) – number of iterations for the refine stage

  • n_points – general number of points per iteration - ignored for survey/refine if n_points_survey/n_points_refine is set

  • n_points_survey (int) – number of points per iteration for the survey stage

  • n_points_refine (int) – number of points per iteration for the refine stage

  • use_survey (bool) – whether to use the points generated during the survey to compute the final integral not recommended due to uncontrolled correlations in error estimates

  • device – pytorch device on which to run

  • verbosity (int) – verbosity level of the integrator

  • trainer_verbosity – verbosity level of the trainer

  • loss (str or function) – loss function used by the trainer

  • flow (str or flow model) – normalizing flow model to use for importance sampling

  • trainer (None or zunis.training.weighted_dataset.weighted_dataset_trainer.BasicStatefulTrainer) – optional argument to provide a full trainer object - overrides any other trainer or model setup argument

  • trainer_options (dict) – dictionary of options to pass to the zunis.training.weighted_dataset.stateful_trainer.StatefulTrainer created if trainer is None

  • flow_options (dict) – dictionary of options to pass to the zunis.models.flows.sequential.repeated_cell.RepeatedCellFlow created by the stateful trainer if trainer is None

Returns

Return type

subclass of zunis.models.flows.sequential.repeated_cell.RepeatedCellFlow