vipr_reflectometry.panpe.predict package

Submodules

vipr_reflectometry.panpe.predict.panpe_predictor module

class vipr_reflectometry.panpe.predict.panpe_predictor.PanpePredictor(**kw: Any)

Bases: PredictorHandler

Predictor handler for PANPE (Physics-Agnostic Neural Posterior Estimation) models.

This predictor integrates PANPE’s probabilistic inference into the VIPR workflow, converting between VIPR’s DataSet format and PANPE’s MeasuredData format.

PANPE performs Bayesian inference to estimate posterior distributions over physical parameters given reflectometry measurements.

class Meta

Bases: object

label = 'panpe_predictor'

class vipr_reflectometry.panpe.predict.panpe_predictor.PanpePredictorParams(*, target_neff: int = 500, max_num_samples: int = 1048576, batch_size_sampling: int = 32768, run_snis: bool = True, prior_bounds: ~typing.Any = <factory>, num_samples_return: int = 1000)

Bases: BaseModel

Parameters for the PANPE predictor.

batch_size_sampling: int

max_num_samples: int

model_config: ClassVar[ConfigDict] = {'extra': 'forbid'}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

num_samples_return: int

prior_bounds: Any

run_snis: bool

target_neff: int

vipr_reflectometry.panpe.predict.panpe_predictor.load(app)

Register the PanpePredictor handler with the application.

Module contents