ScaleWeibull#

class liesel_ptm.ScaleWeibull(value, scale, concentration=0.5, name='', inference=None, bijector=None, role='hyperparam', clip_min='default')[source]#

Bases: Var

A variable with a Weibull prior on its square.

Parameters:
  • value (Any) – Initial value of the variable.

  • concentration (float | Var | Node) – Concentration parameter. (default: 0.5)

  • scale (float | Var | Node) – Scale parameter.

  • name (str) – Name of the variable. (default: '')

  • inference (Any) – Inference type. (default: None)

  • bijector (Bijector | None) – A tensorflow bijector instance. If a bijector is supplied, the variable will be transformed using the bijector. This renders the variable itself weak, meaning that it is a deterministic function of the newly created transformed variable. The prior is transferred to this transformed variable and transformed according to the change-of-variables theorem. (default: None)

  • clip_min (Union[Literal['default'], float, Any]) – Values very close to zero will be soft-clipped to this value to avoid numerical instability. For 32-bit floats, we use a default of jnp.sqrt(jnp.exp(-9.0)); for 64-bit floats we use jnp.sqrt(jnp.exp(-11.0)). (default: 'default')

variance_param#

The internal variance parameter (square of the reported scale). This is an lsl.Var created as the latent variance parameter.

bijector#

The bijector passed to the constructor (or None). When present, the bijector was applied to the variance parameter and the public variable becomes a deterministic transform of that parameter.

Methods

Attributes