Ctrl+K

LuasPyMC

Contents

LuasPyMC#

luas.pymc_ext.LuasPyMC(name, gp=None, var_dict=None, Y=None, likelihood_fn=None, jit=True)[source]#

PyMC extension which can by used with a luas.GPClass.GP object for log-likelihood calculations.

Parameters:

  • name (str) – Name of observable storing likelihood values e.g. “log_like”.

  • gp (object) – The GP object used for log likelihood calculations.

  • var_dict (PyTree) – A PyTree of parameter values for calculating the log likelihood.

  • Y (JAXArray) – Data values being fit.

  • likelihood_fn (Callable, optional) – Can specify a different log likelihood function other than the default of GP.logP_stored. Needs to take the same inputs and give the same order of outputs as GP.logP_stored.

  • jit (bool, optional) – Whether to jit compile the likelihood function, as PyMC does not require the log likelihood to be jit compiled. Defaults to True.

Contents