malt.models.regressor.HeteroschedasticGaussianLikelihood

class malt.models.regressor.HeteroschedasticGaussianLikelihood[source]

Bases: malt.models.regressor.NeuralNetworkLikelihood

Models a heteroschedastic Gaussian likelihood to be used with neural network regressors. (Admits unconstrained parameters.)

Variables

2 (in_features =) –

Parameters

  • mu (torch.Tensor)

  • log_sigma (torch.Tensor)

Examples

>>> likelihood = HeteroschedasticGaussianLikelihood()
>>> posterior = likelihood(torch.tensor(0.0), torch.tensor(0.0))
__init__()

Methods

__init__()

Attributes

in_features