API Reference

This is the reference for the public API. For a more guided introduction, please refer to the Examples section.

Modules

adapters	A collection of Adapter transforms, which tell BayesFlow how to interpret your Simulator output and plug it into neural networks for training and inference.
approximators	A collection of Approximators, which embody the inference task and the neural network components used to perform it.
datasets	A collection of keras.utils.PyDatasets, which wrap your data-generating process (i.e., your Simulator) and thus determine the effective training strategy (e.g., online or offline).
diagnostics	A collection of plotting utilities and metrics for evaluating trained Workflows.
distributions	A collection of Distributions, which represent the latent space for InferenceNetworks or the summary space of SummaryNetworks.
experimental	Unstable or largely untested networks, proceed with caution.
links	Activation functions linking network output to estimates with architecturally enforced properties.
metrics	A collection of keras.Metrics for evaluating the performance of models.
networks	A rich collection of neural network architectures for use in Approximators.
scores	Scoring rules for point estimation.
simulators	This module provides Simulator, make_simulator(), and several other kinds of Simulator implementations, which serve as generative mathematical models, or data generating processes, with their primary function being to sample data.
types	Custom types used for type annotations.
utils	A collection of utility functions, mostly used for internal purposes.
workflows	Provides BasicWorkflow, a high-level interface for working on typical BayesFlow applications without having to worry about the internals.
wrappers