bilby.hyper.model.Model

class bilby.hyper.model.Model(model_functions=None, cache=True)

Bases: object

Population model that combines a set of factorizable models.

This should take population parameters and return the probability.

\[p(\theta | \Lambda) = \prod_{i} p_{i}(\theta | \Lambda)\]

__init__(model_functions=None, cache=True)

Parameters:

model_functions: list

List of callables to compute the probability. If this includes classes, the __call__: method should return the probability. The requires variables are chosen at run time based on either inspection or querying a variable_names attribute.

cache: bool

Whether to cache the value returned by the model functions, default=:code:True. The caching only looks at the parameters not the data, so should be used with caution. The caching also breaks jax JIT compilation.

__call__(*args, **kwargs)

Call self as a function.

Methods

__init__([model_functions, cache])
prob(data, **kwargs)	Compute the total population probability for the provided data given the keyword arguments.

prob(data, **kwargs)

Compute the total population probability for the provided data given the keyword arguments.

Parameters:

data: dict

Dictionary containing the points at which to evaluate the population model.

kwargs: dict

The population parameters. These cannot include any of ["dataset", "data", "self", "cls"] unless the variable_names attribute is available for the relevant model.