pymc.Inference#

class pymc.Inference(op, approx, tf, **kwargs)[source]#

Base class for Variational Inference.

Communicates Operator, Approximation and Test Function to build Objective Function

Parameters:
opOperator class #:class:~pymc.variational.operators
approxApproximation class or instance #:class:~pymc.variational.approximations
tfTestFunction instance #?
modelModel

PyMC Model

kwargskwargs passed to Operator #:class:~pymc.variational.operators, optional

Methods

Inference.__init__(op, approx, tf, **kwargs)

Inference.fit([n, score, callbacks, ...])

Perform Operator Variational Inference.

Inference.refine(n[, progressbar, ...])

Refine the solution using the last compiled step function.

Inference.run_profiling([n, score])

Attributes

approx