template<typename... Args>
class ROOT::Experimental::Internal::ML::RChunkLoader< Args >

Building and loading the chunks from the blocks and chunks constructed in RChunkConstructor.

In this class the blocks are stiches together to form chunks that are loaded into memory. The blocks used to create each chunk comes from different parts of the dataset. This is achieved by shuffling the blocks before distributing them into chunks. The purpose of this process is to reduce bias during machine learning training by ensuring that the data is well mixed. The dataset is also spit into training and validation sets with the user-defined validation split fraction.

Definition at line 112 of file RChunkLoader.hxx.

Public Member Functions
	RChunkLoader (ROOT::RDF::RNode &rdf, const std::size_t chunkSize, const std::size_t blockSize, const float validationSplit, const std::vector< std::string > &cols, const std::vector< std::size_t > &vecSizes={}, const float vecPadding=0.0, bool shuffle=true, const std::size_t setSeed=0)

void	CheckIfOverlap (RFlat2DMatrix &Tensor1, RFlat2DMatrix &Tensor2)

void	CheckIfUnique (RFlat2DMatrix &Tensor)

void	CreateTrainingChunksIntervals ()
	Create training chunks consisiting of block intervals of different types.

void	CreateValidationChunksIntervals ()
	Create training chunks consisiting of block intervals of different types.

std::size_t	GetNumTrainingChunks ()

std::size_t	GetNumTrainingEntries ()

std::size_t	GetNumValidationChunks ()

std::size_t	GetNumValidationEntries ()

std::vector< std::size_t >	GetTrainingChunkSizes ()

std::vector< std::size_t >	GetValidationChunkSizes ()

void	LoadTrainingChunk (RFlat2DMatrix &TrainChunkTensor, std::size_t chunk)
	Load the nth chunk from the training dataset into a tensor.

void	LoadValidationChunk (RFlat2DMatrix &ValidationChunkTensor, std::size_t chunk)
	Load the nth chunk from the validation dataset into a tensor.

void	ResetDataframe ()

void	SplitDataset ()
	Distribute the blocks into training and validation datasets.

Private Attributes
ROOT::RDF::RNode &	f_rdf

std::size_t	fBlockSize

std::size_t	fChunkSize

std::vector< std::string >	fCols

ROOT::RDF::RResultPtr< std::vector< ULong64_t > >	fEntries

bool	fNotFiltered

std::size_t	fNumChunkCols

std::size_t	fNumCols

std::size_t	fNumEntries

std::size_t	fNumTrainEntries

std::size_t	fNumValidationEntries

std::size_t	fSetSeed

bool	fShuffle

std::size_t	fSumVecSizes

std::unique_ptr< RFlat2DMatrixOperators >	fTensorOperators

std::unique_ptr< RChunkConstructor >	fTraining

std::unique_ptr< RChunkConstructor >	fValidation

float	fValidationSplit

std::size_t	fVecPadding

std::vector< std::size_t >	fVecSizes

#include <ROOT/ML/RChunkLoader.hxx>

Constructor & Destructor Documentation

◆ RChunkLoader()

template<typename... Args>

ROOT::Experimental::Internal::ML::RChunkLoader< Args >::RChunkLoader	(	ROOT::RDF::RNode &	rdf,
		const std::size_t	chunkSize,
		const std::size_t	blockSize,
		const float	validationSplit,
		const std::vector< std::string > &	cols,
		const std::vector< std::size_t > &	vecSizes = {},
		const float	vecPadding = 0.0,
		bool	shuffle = true,
		const std::size_t	setSeed = 0 )

inline

Definition at line 142 of file RChunkLoader.hxx.

Member Function Documentation

◆ CheckIfOverlap()

template<typename... Args>

void ROOT::Experimental::Internal::ML::RChunkLoader< Args >::CheckIfOverlap	(	RFlat2DMatrix &	Tensor1,
		RFlat2DMatrix &	Tensor2 )

inline

Definition at line 448 of file RChunkLoader.hxx.

◆ CheckIfUnique()

template<typename... Args>

void ROOT::Experimental::Internal::ML::RChunkLoader< Args >::CheckIfUnique ( RFlat2DMatrix & Tensor )

inline

Definition at line 440 of file RChunkLoader.hxx.

◆ CreateTrainingChunksIntervals()

template<typename... Args>

void ROOT::Experimental::Internal::ML::RChunkLoader< Args >::CreateTrainingChunksIntervals ( )

inline

Create training chunks consisiting of block intervals of different types.

Definition at line 255 of file RChunkLoader.hxx.

◆ CreateValidationChunksIntervals()

template<typename... Args>

void ROOT::Experimental::Internal::ML::RChunkLoader< Args >::CreateValidationChunksIntervals ( )

inline

Create training chunks consisiting of block intervals of different types.

Definition at line 295 of file RChunkLoader.hxx.

◆ GetNumTrainingChunks()

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::GetNumTrainingChunks ( )

inline

Definition at line 471 of file RChunkLoader.hxx.

◆ GetNumTrainingEntries()

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::GetNumTrainingEntries ( )

inline

Definition at line 437 of file RChunkLoader.hxx.

◆ GetNumValidationChunks()

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::GetNumValidationChunks ( )

inline

Definition at line 473 of file RChunkLoader.hxx.

◆ GetNumValidationEntries()

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::GetNumValidationEntries ( )

inline

Definition at line 438 of file RChunkLoader.hxx.

◆ GetTrainingChunkSizes()

template<typename... Args>

std::vector< std::size_t > ROOT::Experimental::Internal::ML::RChunkLoader< Args >::GetTrainingChunkSizes ( )

inline

Definition at line 434 of file RChunkLoader.hxx.

◆ GetValidationChunkSizes()

template<typename... Args>

std::vector< std::size_t > ROOT::Experimental::Internal::ML::RChunkLoader< Args >::GetValidationChunkSizes ( )

inline

Definition at line 435 of file RChunkLoader.hxx.

◆ LoadTrainingChunk()

template<typename... Args>

void ROOT::Experimental::Internal::ML::RChunkLoader< Args >::LoadTrainingChunk	(	RFlat2DMatrix &	TrainChunkTensor,
		std::size_t	chunk )

inline

Load the nth chunk from the training dataset into a tensor.

Parameters

[in]	TrainChunkTensor	RTensor for the training chunk
[in]	chunk	Index of the chunk in the dataset

Definition at line 333 of file RChunkLoader.hxx.

◆ LoadValidationChunk()

template<typename... Args>

void ROOT::Experimental::Internal::ML::RChunkLoader< Args >::LoadValidationChunk	(	RFlat2DMatrix &	ValidationChunkTensor,
		std::size_t	chunk )

inline

Load the nth chunk from the validation dataset into a tensor.

Parameters

[in]	ValidationChunkTensor	RTensor for the validation chunk
[in]	chunk	Index of the chunk in the dataset

Definition at line 385 of file RChunkLoader.hxx.

◆ ResetDataframe()

template<typename... Args>

void ROOT::Experimental::Internal::ML::RChunkLoader< Args >::ResetDataframe ( )

inline

Definition at line 432 of file RChunkLoader.hxx.

◆ SplitDataset()

template<typename... Args>

void ROOT::Experimental::Internal::ML::RChunkLoader< Args >::SplitDataset ( )

inline

Distribute the blocks into training and validation datasets.

Definition at line 180 of file RChunkLoader.hxx.

Member Data Documentation

◆ f_rdf

template<typename... Args>

ROOT::RDF::RNode& ROOT::Experimental::Internal::ML::RChunkLoader< Args >::f_rdf

private

Definition at line 128 of file RChunkLoader.hxx.

◆ fBlockSize

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fBlockSize

private

Definition at line 116 of file RChunkLoader.hxx.

◆ fChunkSize

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fChunkSize

private

Definition at line 115 of file RChunkLoader.hxx.

◆ fCols

template<typename... Args>

std::vector<std::string> ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fCols

private

Definition at line 129 of file RChunkLoader.hxx.

◆ fEntries

template<typename... Args>

ROOT::RDF::RResultPtr<std::vector<ULong64_t> > ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fEntries

private

Definition at line 136 of file RChunkLoader.hxx.

◆ fNotFiltered

template<typename... Args>

bool ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fNotFiltered

private

Definition at line 133 of file RChunkLoader.hxx.

◆ fNumChunkCols

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fNumChunkCols

private

Definition at line 122 of file RChunkLoader.hxx.

◆ fNumCols

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fNumCols

private

Definition at line 130 of file RChunkLoader.hxx.

◆ fNumEntries

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fNumEntries

private

Definition at line 114 of file RChunkLoader.hxx.

◆ fNumTrainEntries

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fNumTrainEntries

private

Definition at line 124 of file RChunkLoader.hxx.

◆ fNumValidationEntries

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fNumValidationEntries

private

Definition at line 125 of file RChunkLoader.hxx.

◆ fSetSeed

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fSetSeed

private

Definition at line 131 of file RChunkLoader.hxx.

◆ fShuffle

template<typename... Args>

bool ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fShuffle

private

Definition at line 134 of file RChunkLoader.hxx.

◆ fSumVecSizes

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fSumVecSizes

private

Definition at line 120 of file RChunkLoader.hxx.

◆ fTensorOperators

template<typename... Args>

std::unique_ptr<RFlat2DMatrixOperators> ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fTensorOperators

private

Definition at line 126 of file RChunkLoader.hxx.

◆ fTraining

template<typename... Args>

std::unique_ptr<RChunkConstructor> ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fTraining

private

Definition at line 138 of file RChunkLoader.hxx.

◆ fValidation

template<typename... Args>

std::unique_ptr<RChunkConstructor> ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fValidation

private

Definition at line 139 of file RChunkLoader.hxx.

◆ fValidationSplit

template<typename... Args>

float ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fValidationSplit

private

Definition at line 117 of file RChunkLoader.hxx.

◆ fVecPadding

template<typename... Args>

std::size_t ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fVecPadding

private

Definition at line 121 of file RChunkLoader.hxx.

◆ fVecSizes

template<typename... Args>

std::vector<std::size_t> ROOT::Experimental::Internal::ML::RChunkLoader< Args >::fVecSizes

private

Definition at line 119 of file RChunkLoader.hxx.

tree/ml/inc/ROOT/ML/RChunkLoader.hxx

Public Member Functions

Private Attributes

Constructor & Destructor Documentation

◆ RChunkLoader()

Member Function Documentation

◆ CheckIfOverlap()

◆ CheckIfUnique()

◆ CreateTrainingChunksIntervals()

◆ CreateValidationChunksIntervals()

◆ GetNumTrainingChunks()

◆ GetNumTrainingEntries()

◆ GetNumValidationChunks()

◆ GetNumValidationEntries()

◆ GetTrainingChunkSizes()

◆ GetValidationChunkSizes()

◆ LoadTrainingChunk()

◆ LoadValidationChunk()

◆ ResetDataframe()

◆ SplitDataset()

Member Data Documentation

◆ f_rdf

◆ fBlockSize

◆ fChunkSize

◆ fCols

◆ fEntries

◆ fNotFiltered

◆ fNumChunkCols

◆ fNumCols

◆ fNumEntries

◆ fNumTrainEntries

◆ fNumValidationEntries

◆ fSetSeed

◆ fShuffle

◆ fSumVecSizes

◆ fTensorOperators

◆ fTraining

◆ fValidation

◆ fValidationSplit

◆ fVecPadding

◆ fVecSizes