Welcome to redbnn’s documentation!¶
Reduced Bayesian Neural Networks are deterministic Neural Networks with a Bayesian subset of weights. redBNN class computes a MAP estimate of the entire NN and then performs Bayesian inference on a chosen layer (--reduction=layers) or block (--reduction=blocks).
This library is built upon pyro and torchvision: redbnn loads any pre-trained architecture from torchvision library and trains a deterministic Neural Network (baseNN) or a reduced Bayesian Neural Network (redBNN) using Stochastic Variational Inference (SVI) or Hamiltonian Monte Carlo (HMC) from pyro library.
