API Reference

BDEEstimator

ml.models.BDEEstimator([model_class, ...])

SKlearn-compatible implementation of a BDE estimator.

GaussianNLLLoss

ml.loss.GaussianNLLLoss([epsilon, ...])

Gaussian negative log likelihood loss.

Utilities

utils.discovery.all_estimators([type_filter])

Get a list of all estimators from bde.

ML Module

Machine Learning Module for Bayesian Deep Ensembles (BDE).

The bde.ml module provides the core machine learning components required for building and training Bayesian Neural Networks within the Bayesian Deep Ensembles (BDE) framework.

This module includes submodules for defining loss functions, neural network models, and training procedures, enabling flexible and robust implementation of BDE models.

Submodules

• datasets: Handles data and dataset management.

• loss: Contains loss functions implementations and loss function related utilities.

• models: Defines the neural network architectures supported by the BDE framework.

• training: Implements the training algorithms and routines used for model optimization.

Example Usage

# TODO: Provide examples

>>> # TODO: Provide an example
>>>
>>>
>>>

Utils Module

Utility functions and configuration for the Bayesian Deep Ensembles (BDE) framework.

This package includes various utility functions and configurations that support the operation of the BDE framework. It provides reusable components such as configuration management tools.

Modules

• configs: Contains configuration management utilities for the BDE framework.

• utils: Contains general utility functions used by other modules.