GeneratorSampler

class aixd.mlmodel.generation.sampling.GeneratorSampler(sampling_type: str = 'kde', datamodule: DataModule | None = None, engine: str = 'sobol')

Bases: object

The Generator class needs to sampled from the outputML, in order to provide these values as entry to the decoder. For that, it uses this specific sampler for the generator.

Parameters:

• sampling_type (str, optional, default=”kde”) – The type of sampling to be used for generation. The options are: bayesian, and kde. kde is the default as it is faster, using a KDE to fitted on the training data. Bayesian is more time consuming, as it uses an objective to perform Bayesian optimization.

• datamodule (DataModule optional, default=None) – The LightningDataModule that handles data loading and preprocessing.

• engine (str, optional, default=”sobol”) – Engine used by the sampler to generate samples. The options are: “bayesian”, “sobol”, “lhc”, and “random”. Run the method summary_sampling_types to obtain more information.

Methods

generate	The Generator class uses the instance provided, and now call generate to obtain the samples required to be passed to the decoder.
generate_z	The Generator class uses the instance provided, and now call generate to obtain the samples required to be passed to the decoder.
summary_sampling_types	Just printing some extra information about the sampling types and the engines