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

PositiveReal#

class gpjax.parameters.PositiveReal(value)[source]#

Bases: AbstractUnwrappable

Strictly positive parameter.

Stored unconstrained via the inverse of the softplus transform; unwrap() applies softplus to recover the constrained value.

Expand for references to gpjax.parameters.PositiveReal

Backend Module Design

Kernel Guide

unwrap()[source]#: Returns the unwrapped pytree, assuming no wrapped subnodes exist.