global_optimisation

jaxns.experimental.global_optimisation

Module Contents

class GlobalOptimisationState[source]

Bases: NamedTuple

key: jaxns.internals.types.PRNGKey[source]

samples: jaxns.nested_samplers.common.types.SampleCollection[source]

num_samples: jaxns.internals.types.IntArray[source]

relative_spread: jaxns.internals.types.FloatArray[source]

absolute_spread: jaxns.internals.types.FloatArray[source]

num_likelihood_evaluations: jaxns.internals.types.IntArray[source]

class GlobalOptimisationResults[source]

Bases: NamedTuple

U_solution: jaxns.internals.types.UType[source]

X_solution: jaxns.internals.types.XType[source]

solution: jaxns.internals.types.LikelihoodInputType[source]

log_L_solution: jaxns.internals.types.FloatArray[source]

log_L_progress: jaxns.internals.types.FloatArray[source]

num_likelihood_evaluations: jaxns.internals.types.IntArray[source]

num_samples: jaxns.internals.types.IntArray[source]

termination_reason: jaxns.internals.types.IntArray[source]

relative_spread: jaxns.internals.types.FloatArray[source]

absolute_spread: jaxns.internals.types.FloatArray[source]

class GlobalOptimisationTerminationCondition[source]

Bases: NamedTuple

max_likelihood_evaluations: jaxns.internals.types.IntArray | None = None[source]

log_likelihood_contour: jaxns.internals.types.FloatArray | None = None[source]

rtol: jaxns.internals.types.FloatArray | None = None[source]

atol: jaxns.internals.types.FloatArray | None = None[source]

min_efficiency: jaxns.internals.types.FloatArray | None = None[source]

class SimpleGlobalOptimisation[source]

Simple global optimisation leveraging building blocks of nested sampling.

sampler: jaxns.samplers.abc.AbstractSampler[source]

num_search_chains: int[source]

model: jaxns.framework.bases.BaseAbstractModel[source]

shell_frac: float = 0.5[source]

devices: jaxlib.xla_client.Device | None = None[source]

verbose: bool = False[source]

__post_init__()[source]