Smoke render: optimisation and MCMC are shrunk for CI speed, so figures and diagnostics are not publication-fidelity. The production docs are built with full-length inference.

GPJax

SoftplusTransform#

class gpjax.likelihoods.SoftplusTransform(num_points=20)[source]#

Bases: AbstractNoiseTransform

Softplus noise transformation.

Expand for references to gpjax.likelihoods.SoftplusTransform

Heteroscedastic Inference / Background

Parameters:: num_points (int)

moments(mean, variance)[source]#

Compute the moments of the transformed noise signal.

Parameters:

mean (Float[Array, ...])
variance (Float[Array, ...])

Return type:

NoiseMoments