LossIndexBackPropagationLM Struct Reference

A loss index composed of several terms, this structure represent the First Order for this function. More...

#include <loss_index.h>

Public Member Functions

 LossIndexBackPropagationLM ()
 Default constructor. More...
 
 LossIndexBackPropagationLM (const Index &new_batch_samples_number, LossIndex *new_loss_index_pointer)
 
void set (const Index &new_batch_samples_number, LossIndex *new_loss_index_pointer)
 
void print () const
 

Public Attributes

LossIndexloss_index_pointer = nullptr
 
Index batch_samples_number = 0
 
type error = type(0)
 
type loss = type(0)
 
Tensor< type, 1 > parameters
 
NeuralNetworkBackPropagationLM neural_network
 
Tensor< type, 2 > errors
 
Tensor< type, 1 > squared_errors
 
Tensor< type, 2 > squared_errors_jacobian
 
Tensor< type, 1 > gradient
 
Tensor< type, 2 > hessian
 
Tensor< type, 1 > regularization_gradient
 
Tensor< type, 2 > regularization_hessian
 

Detailed Description

A loss index composed of several terms, this structure represent the First Order for this function.

This structure contains second order information about the loss function (loss, gradient and Hessian). Set of loss value, gradient vector and Hessian matrix of the loss index. A method returning this structure might be implemented more efficiently than the loss, gradient and Hessian methods separately.

Definition at line 382 of file loss_index.h.

Constructor & Destructor Documentation

◆ LossIndexBackPropagationLM() [1/2]

Default constructor.

Definition at line 386 of file loss_index.h.

◆ LossIndexBackPropagationLM() [2/2]

LossIndexBackPropagationLM ( const Index &  new_batch_samples_number,
LossIndex new_loss_index_pointer 
)
inlineexplicit

Definition at line 388 of file loss_index.h.

Member Function Documentation

◆ print()

void print ( ) const
inline

Definition at line 432 of file loss_index.h.

◆ set()

void set ( const Index &  new_batch_samples_number,
LossIndex new_loss_index_pointer 
)
inline

Definition at line 395 of file loss_index.h.

Member Data Documentation

◆ batch_samples_number

Index batch_samples_number = 0

Definition at line 460 of file loss_index.h.

◆ error

type error = type(0)

Definition at line 462 of file loss_index.h.

◆ errors

Tensor<type, 2> errors

Definition at line 469 of file loss_index.h.

◆ gradient

Tensor<type, 1> gradient

Definition at line 473 of file loss_index.h.

◆ hessian

Tensor<type, 2> hessian

Definition at line 474 of file loss_index.h.

◆ loss

type loss = type(0)

Definition at line 463 of file loss_index.h.

◆ loss_index_pointer

LossIndex* loss_index_pointer = nullptr

Definition at line 458 of file loss_index.h.

◆ neural_network

Definition at line 467 of file loss_index.h.

◆ parameters

Tensor<type, 1> parameters

Definition at line 465 of file loss_index.h.

◆ regularization_gradient

Tensor<type, 1> regularization_gradient

Definition at line 476 of file loss_index.h.

◆ regularization_hessian

Tensor<type, 2> regularization_hessian

Definition at line 477 of file loss_index.h.

◆ squared_errors

Tensor<type, 1> squared_errors

Definition at line 470 of file loss_index.h.

◆ squared_errors_jacobian

Tensor<type, 2> squared_errors_jacobian

Definition at line 471 of file loss_index.h.


The documentation for this struct was generated from the following file: