options

Documentation

Overview

Package options contains implementation of types and constants related to the regression options.

Index

• type ConvergenceType
• type GradientDescentVariant
• type Options
  • func WithAutomaticConvergence(lr float64, gdv GradientDescentVariant, t float64) Options
  • func WithIterativeConvergence(lr float64, gdv GradientDescentVariant, i uint) Options

Types

type ConvergenceType

type ConvergenceType int

ConvergenceType identifies a convergence type.

const (
	Iterative ConvergenceType = iota + 1
	Automatic
)

type GradientDescentVariant

type GradientDescentVariant int

GradientDescentVariant identifies a gradient descent variant.

const (
	Batch GradientDescentVariant = iota + 1
	Stochastic
)

type Options

type Options struct {
	LearningRate           float64
	GradientDescentVariant GradientDescentVariant
	ConvergenceType        ConvergenceType
	ConvergenceIndicator   float64
}

Options contains training options for a regression with gradient descent.

func WithAutomaticConvergence

func WithAutomaticConvergence(lr float64, gdv GradientDescentVariant, t float64) Options

WithAutomaticConvergence returns new training options with an automatic convergence indicator.

func WithIterativeConvergence

func WithIterativeConvergence(lr float64, gdv GradientDescentVariant, i uint) Options

WithIterativeConvergence returns new training options with an iterative convergence indicator.