Structure containing the optimization options for hyperparameter-tuning. More...

#include <GaussianProcess.h>

Public Member Functions
	GPOptimizerOptions (const unsigned int show_every_nth_iteration=0, const unsigned int num_iter=1000, const unsigned int batch_size=0, const Real learning_rate=1e-3, const Real b1=0.9, const Real b2=0.999, const Real eps=1e-7, const Real lambda=1e-4, const OptimizerType optimizer_type=OptimizerType::Adam)
	Construct a new GPOptimizerOptions object using input parameters that will control the optimization. More...

Public Attributes
const unsigned int	show_every_nth_iteration = 0
	Switch to enable verbose output for parameter tuning at every n-th iteration. More...

const unsigned int	num_iter = 1000
	The number of iterations for Adam optimizer. More...

const unsigned int	batch_size = 0
	The batch size for Adam optimizer. More...

const Real	learning_rate = 1e-3
	The learning rate for Adam optimizer. More...

const Real	b1 = 0.9
	Tuning parameter from the paper. More...

const Real	b2 = 0.999
	Tuning parameter from the paper. More...

const Real	eps = 1e-7
	Tuning parameter from the paper. More...

const Real	lambda = 1e-4
	Legacy MOOSE shrink parameter. More...

const OptimizerType	optimizer_type = OptimizerType::Adam
	Adam optimizer mode to use. More...

Detailed Description

Structure containing the optimization options for hyperparameter-tuning.

Definition at line 57 of file GaussianProcess.h.

Constructor & Destructor Documentation

◆ GPOptimizerOptions()

StochasticTools::GaussianProcess::GPOptimizerOptions::GPOptimizerOptions	(	const unsigned int	show_every_nth_iteration = `0`,
		const unsigned int	num_iter = `1000`,
		const unsigned int	batch_size = `0`,
		const Real	learning_rate = `1e-3`,
		const Real	b1 = `0.9`,
		const Real	b2 = `0.999`,
		const Real	eps = `1e-7`,
		const Real	lambda = `1e-4`,
		const OptimizerType	optimizer_type = `OptimizerType::Adam`
	)

Construct a new GPOptimizerOptions object using input parameters that will control the optimization.

Parameters

show_every_nth_iteration	To show the loss value at every n-th iteration, if set to 0, nothing is displayed
num_iter	The number of iterations we want in the optimization of the GP
batch_size	The number of samples in each batch
learning_rate	The learning rate for parameter updates
b1	Tuning constant for the Adam algorithm
b2	Tuning constant for the Adam algorithm
eps	Tuning constant for the Adam algorithm
lambda	Legacy MOOSE shrink constant for the Adam algorithm
optimizer_type	The Adam optimizer mode to use

Definition at line 108 of file GaussianProcess.C.

   : show_every_nth_iteration(show_every_nth_iteration),
     num_iter(num_iter),
     batch_size(batch_size),
     learning_rate(learning_rate),
     b1(b1),
     b2(b2),
     eps(eps),
     lambda(lambda),
     optimizer_type(optimizer_type)
 {
 }