distributions (pmrf.distributions)

Probability distribution classes and factories.

Can be used for random parameters in pmrf.parameters.

These are (mostly) re-exports from the distreqx library, with the goal of the covering the most common applications. For more advanced use-cases, simply use distreqx directly instead.

Classes

AbstractDistribution()	Base class for all distreqx distributions.
Normal(loc, scale)	Normal distribution with location loc and scale parameters.
Uniform(low, high)	Uniform distribution with low and high parameters.
Gamma(concentration, rate)	Gamma distribution with parameters concentration and rate.
Transformed(distribution, bijector)	Distribution of a random variable transformed by a bijective function.
LogNormal(loc, scale)	LogNormal distribution parameterized by loc and scale of the underlying Normal.
TruncatedNormal(loc, scale, low, high)	Truncated Normal distribution with loc, scale, low, and high parameters.
Joint(distributions)	Joint distribution over a PyTree of statistically independent distributions.

Functions

CenteredUniform(center, half_width)	Create a Uniform distribution defined by center and width.
RelativeNormal(mean, pct_std)	Create a Normal distribution defined by relative standard deviation.
RelativeTruncatedNormal(mean, pct_std[, ...])	Create a symmetric Truncated Normal distribution defined by relative standard deviation and truncated at a specified number of standard deviations from the mean.
truncate(dist, new_lower, new_upper)	Attempts to return a truncated version of the distribution.