WorkGenerator< T > Class Template Reference

#include <WorkGenerator.h>

Inheritance diagram for WorkGenerator< T >:

Public Member Functions
	WorkGenerator ()=default
	WorkGenerator (WorkGenerator &&) noexcept=default
	WorkGenerator (const WorkGenerator &)=default
WorkGenerator &	operator= (WorkGenerator &&) noexcept=default
WorkGenerator &	operator= (const WorkGenerator &)=default
virtual	~WorkGenerator ()=default
std::pair< std::size_t, T >	next (std::size_t n)
	Generate the next n batches of work.
std::size_t	offset () const
	Return the current offset, i.e., the number of batches that have been generated.
virtual bool	has_more () const =0
	Whether the generator has more work to generate.

Protected Member Functions
virtual std::pair< std::size_t, T >	generate (std::size_t n)=0
	Generate the next n batches of work.

Constructor & Destructor Documentation

WorkGenerator() [1/3]

template<typename T >

WorkGenerator ( )

default

WorkGenerator() [2/3]

template<typename T >

WorkGenerator ( WorkGenerator< T > && )

defaultnoexcept

WorkGenerator() [3/3]

template<typename T >

WorkGenerator ( const WorkGenerator< T > & )

default

~WorkGenerator()

template<typename T >

virtual ~WorkGenerator ( )

virtualdefault

Member Function Documentation

generate()

template<typename T >

virtual std::pair< std::size_t, T > generate ( std::size_t n )

protectedpure virtual

Generate the next n batches of work.

Note that in the case of insufficient remaining work, it is possible that the number of batches generated is less than n.

Parameters

n	Number of batches to generate

Returns

std::pair<std::size_t, T> Number of batches generated (m) and the next m batches of work

Implemented in SliceGenerator, TensorLoader, and ValueMapLoader.

has_more()

template<typename T >

virtual bool has_more ( ) const

pure virtual

Whether the generator has more work to generate.

Implemented in FixedSizeWorkGenerator< T >, FixedSizeWorkGenerator< indexing::Slice >, FixedSizeWorkGenerator< Tensor >, and FixedSizeWorkGenerator< ValueMap >.

next()

template<typename T >

std::pair< std::size_t, T > next ( std::size_t n )

inline

Generate the next n batches of work.

Note that in the case of insufficient remaining work, it is possible that the number of batches generated is less than n.

This is the public interface to the generator. Derived classes should implement WorkGenerator::generate.

Parameters

n	Number of batches to generator

Returns

std::pair<std::size_t, T> Number of batches generated (m) and the next m batches of work

offset()

template<typename T >

std::size_t offset ( ) const

inline

Return the current offset, i.e., the number of batches that have been generated.

operator=() [1/2]

template<typename T >

WorkGenerator & operator= ( const WorkGenerator< T > & )

default

operator=() [2/2]

template<typename T >

WorkGenerator & operator= ( WorkGenerator< T > && )

defaultnoexcept