redbnn.bayesian_inference package

Submodules

redbnn.bayesian_inference.hmc module

redbnn.bayesian_inference.hmc.forward(redbnn, inputs, n_samples, sample_idxs=None, softmax=True)

Forward pass of the inputs through the network using the chosen number of samples.

Parameters

  • inputs (torch.tensor) – Input images.

  • n_samples (int, optional) – Number of samples drawn during the evaluation.

  • samples_idxs (list, optional) – Random seeds used for drawing samples. If samples_idxs is None it is defined as the range of integers from 0 to the maximum number of samples.

  • softmax (bool, optional) – If True computes the softmax of each output tensor.

Returns

Output predictions

Return type

(torch.Tensor)

redbnn.bayesian_inference.hmc.load(redbnn, savedir, filename, hmc_samples)

Loads the learned parameters.

Parameters

  • savedir (str) – Output directory.

  • filename (str) – Filename.

  • hmc_samples (str) – Number of samples drawn during HMC inference, needed for loading models trained with HMC.

redbnn.bayesian_inference.hmc.model(redbnn, x_data, y_data)

Stochastic function that implements the generative process and is conditioned on the observations.

Parameters

  • x_data (torch.tensor) – Observed data points.

  • y_data (torch.tensor) – Labels of the observed data.

redbnn.bayesian_inference.hmc.save(redbnn, savedir, filename, hmc_samples)

Saves the learned parameters as torch.tensors on the CPU.

Parameters

  • savedir (str) – Output directory.

  • filename (str) – Filename.

  • hmc_samples (str) – Number of samples drawn during HMC inference, needed for saving models trained with HMC.

redbnn.bayesian_inference.hmc.to(device)

Sends pyro parameters to the chosen device.

Parameters

device (str) – Name of the chosen device.

redbnn.bayesian_inference.hmc.train(redbnn, dataloaders, device, n_samples, warmup, is_inception)

Freezes the deterministic parameters and infers the Bayesian paramaters using Hamiltonian Monte Carlo.

Parameters

  • dataloaders (dict) – Dictionary containing training and validation torch dataloaders.

  • device (str) – Device chosen for training.

  • n_samples (int) – Number of Hamiltonian Monte Carlo samples.

  • warmup (int) – Number of Hamiltonian Monte Carlo warmup samples.

redbnn.bayesian_inference.svi module

redbnn.bayesian_inference.svi.forward(redbnn, inputs, n_samples, sample_idxs=None, softmax=True)

Forward pass of the inputs through the network using the chosen number of samples.

Parameters

  • inputs (torch.tensor) – Input images.

  • n_samples (int, optional) – Number of samples drawn during the evaluation.

  • samples_idxs (list, optional) – Random seeds used for drawing samples. If samples_idxs is None it is defined as the range of integers from 0 to the maximum number of samples.

  • softmax (bool, optional) – If True computes the softmax of each output tensor.

Returns

Output predictions

Return type

(torch.Tensor)

redbnn.bayesian_inference.svi.guide(redbnn, x_data, y_data=None)

Variational distribution.

Parameters

  • x_data (torch.tensor) – Input data points.

  • y_data (torch.tensor, optional) – Labels of the input data.

redbnn.bayesian_inference.svi.load(redbnn, savedir, filename)

Loads the learned parameters.

Parameters

  • savedir (str) – Output directory.

  • filename (str) – Filename.

redbnn.bayesian_inference.svi.model(redbnn, x_data, y_data)

Stochastic function that implements the generative process and is conditioned on the observations.

Parameters

  • x_data (torch.tensor) – Observed data points.

  • y_data (torch.tensor) – Labels of the observed data.

redbnn.bayesian_inference.svi.save(redbnn, savedir, filename)

Saves the learned parameters on the CPU.

Parameters

  • savedir (str) – Output directory.

  • filename (str) – Filename.

redbnn.bayesian_inference.svi.to(device)

Sends pyro parameters to the chosen device.

Parameters

device (str) – Name of the chosen device.

redbnn.bayesian_inference.svi.train(redbnn, dataloaders, device, num_iters, is_inception, lr=0.01)

Freezes the deterministic parameters and infers the Bayesian paramaters using the chosen inference method.

Parameters

  • dataloaders (dict) – Dictionary containing training and validation torch dataloaders.

  • device (str) – Device chosen for training.

  • num_iters (int) – Number of iterations for Stochastic Variational Inference.

  • lr (float, optional) – Learning rate for SVI.

Module contents