txeo::TensorShape Class Reference

The shape of a tensor is an ordered collection of dimensions of mathematical vector spaces. More...

#include <TensorShape.h>

Collaboration diagram for txeo::TensorShape:

Public Member Functions
	TensorShape (const TensorShape &shape)
	TensorShape (TensorShape &&shape) noexcept
	~TensorShape ()
TensorShape &	operator= (const TensorShape &shape)
TensorShape &	operator= (TensorShape &&shape) noexcept
bool	operator== (const TensorShape &shape) const
bool	operator!= (const TensorShape &shape) const
	TensorShape (int number_of_axes, size_t dim)
	Constructs a tensor shape with axes having the same dimension.
	TensorShape (const std::vector< size_t > &shape)
	Constructs a tensor shape from a std::vector.
	TensorShape (std::vector< size_t > &&shape)
	Constructs a tensor shape from a std::vector.
	TensorShape (const std::initializer_list< size_t > &shape)
	Constructs a tensor shape from an initializer list.
int	number_of_axes () const noexcept
	Returns the size of the tensor shape.
int	size () const noexcept
	Synonym for txeo::TensorShape::number_of_axes()
int64_t	axis_dim (int axis) const
	Returns the dimension of the specified axis.
const std::vector< size_t > &	stride () const
	Returns the stride of each dimension in the tensor.
std::vector< int64_t >	axes_dims () const noexcept
	Returns the collection of tensor shape dimensions.
bool	is_fully_defined () const noexcept
	Indicates whether the tensor shape has any negative(undefined) dimensions or not.
void	push_axis_back (size_t dim)
	Inserts a dimension after the last axis.
void	insert_axis (int axis, size_t dim)
	Inserts a dimension at the specified axis.
void	remove_axis (int axis)
	Removes a specified axis.
void	remove_all_axes ()
	Removes all axes from this shape, resulting an empty shape.
void	set_dim (int axis, size_t dim)
	Sets a dimension in a specified axis.
size_t	calculate_capacity () const noexcept
	Calculates the number of available tensor elements specified by the tensor shape.
TensorShape	clone () const
	Returns a clone of this tensor.

Friends
template<typename T >
class	Tensor
template<typename T >
class	Predictor
template<typename T >
class	TensorAgg
class	txeo::detail::TensorHelper
std::ostream &	operator<< (std::ostream &os, const TensorShape &shape)

Detailed Description

The shape of a tensor is an ordered collection of dimensions of mathematical vector spaces.

Each position of the tensor shape is an "axis", labeled starting from zero, and the value in this position is a "dimension". An empty tensor shape is the shape of a scalar value. In some some conditions, a negative dimension represents an undefined dimension, but this attribution is reserved to TensorFlow C++.

Definition at line 30 of file TensorShape.h.

Constructor & Destructor Documentation

TensorShape() [1/6]

txeo::TensorShape::TensorShape

(

const TensorShape &

shape

)

TensorShape() [2/6]

txeo::TensorShape::TensorShape ( TensorShape &&  shape )

noexcept

~TensorShape()

txeo::TensorShape::~TensorShape

(

)

TensorShape() [3/6]

txeo::TensorShape::TensorShape ( int  number_of_axes, size_t  dim  )

explicit

Constructs a tensor shape with axes having the same dimension.

Parameters

number_of_axes	Number of axes of the tensor
dim	dimension in each axis

Exceptions

txeo::TensorShapeError

Example Usage:

#include <iostream>
#include "txeo/TensorShape.h"
 
int main() {
    txeo::TensorShape shape(3, 5); // A tensor shape with 3 dimensions, each of size 5
 
    std::cout << "TensorShape created: " << shape << std::endl;
    return 0;
}

TensorShape() [4/6]

txeo::TensorShape::TensorShape ( const std::vector< size_t > &  shape )

explicit

Constructs a tensor shape from a std::vector.

Parameters

shape

vector of dimensions

Example Usage:

#include <iostream>
#include "txeo/TensorShape.h"
 
int main() {
    std::vector<size_t> dims = {3, 4, 5};
    txeo::TensorShape shape(dims); // Creates a 3D shape with dimensions 3x4x5
 
    std::cout << "TensorShape created: " << shape << std::endl;
    return 0;
}

TensorShape() [5/6]

txeo::TensorShape::TensorShape ( std::vector< size_t > &&  shape )

explicit

Constructs a tensor shape from a std::vector.

Parameters

shape

vector of dimensions Example Usage: #include <iostream> #include "txeo/TensorShape.h" int main() { txeo::TensorShape shape({3, 4, 5}); // Creates a 3D shape with dimensions 3x4x5 std::cout << "TensorShape created: " << shape << std::endl; return 0; }

TensorShape() [6/6]

txeo::TensorShape::TensorShape ( const std::initializer_list< size_t > &  shape )

inlineexplicit

Constructs a tensor shape from an initializer list.

Parameters

shape

vector of dimensions Example Usage: #include <iostream> #include "txeo/TensorShape.h" int main() { txeo::TensorShape shape({3, 4, 5}); // Creates a 3D shape with dimensions 3x4x5 std::cout << "TensorShape created: " << shape << std::endl; return 0; }

Definition at line 123 of file TensorShape.h.

        : TensorShape(std::vector<size_t>(shape)) {}

Member Function Documentation

axes_dims()

std::vector< int64_t > txeo::TensorShape::axes_dims ( ) const

noexcept

Returns the collection of tensor shape dimensions.

Returns

std::vector<int64_t>

axis_dim()

int64_t txeo::TensorShape::axis_dim

(

int

axis

)

const

Returns the dimension of the specified axis.

Parameters

axis	Axis whose dimension is to be returned

Returns

int64_t Dimension

Exceptions

txeo::TensorShapeError

calculate_capacity()

size_t txeo::TensorShape::calculate_capacity ( ) const

noexcept

Calculates the number of available tensor elements specified by the tensor shape.

Returns

int64_t Number of tensor elements

clone()

TensorShape txeo::TensorShape::clone

(

)

const

Returns a clone of this tensor.

Returns

TensorShape Clone of this tensor

insert_axis()

void txeo::TensorShape::insert_axis	(	int	axis,
		size_t	dim
	)

Inserts a dimension at the specified axis.

Parameters

axis	Specified axis
dim	Specified dimension

Exceptions

txeo::TensorShapeError

Example Usage:

#include <iostream>
#include "txeo/TensorShape.h"
 
int main() {
    txeo::TensorShape shape({2, 3}); // Initial shape (2,3)
    shape.insert_axis(1, 4); // Inserts a new axis at index 1, resulting in shape (2,4,3)
 
    std::cout << "Updated TensorShape: " << shape << std::endl;
    return 0;
}

is_fully_defined()

bool txeo::TensorShape::is_fully_defined ( ) const

noexcept

Indicates whether the tensor shape has any negative(undefined) dimensions or not.

Returns

true Does not have a negative dimension

false Has a negative dimension

number_of_axes()

int txeo::TensorShape::number_of_axes ( ) const

noexcept

Returns the size of the tensor shape.

Returns

int Number of axes

Here is the caller graph for this function:

operator!=()

bool txeo::TensorShape::operator!=

(

const TensorShape &

shape

)

const

operator=() [1/2]

TensorShape & txeo::TensorShape::operator=

(

const TensorShape &

shape

)

operator=() [2/2]

TensorShape & txeo::TensorShape::operator= ( TensorShape &&  shape )

noexcept

operator==()

bool txeo::TensorShape::operator==

(

const TensorShape &

shape

)

const

push_axis_back()

void txeo::TensorShape::push_axis_back

(

size_t

dim

)

Inserts a dimension after the last axis.

Parameters

dim	Specified dimension

Example Usage:

#include <iostream>
#include "txeo/TensorShape.h"
 
int main() {
    txeo::TensorShape shape({2, 3}); // Initial shape (2,3)
    shape.push_axis_back(4); // Adds a new axis, resulting in shape (2,3,4)
 
    std::cout << "Updated TensorShape: " << shape << std::endl;
    return 0;
}

remove_all_axes()

void txeo::TensorShape::remove_all_axes

(

)

Removes all axes from this shape, resulting an empty shape.

remove_axis()

void txeo::TensorShape::remove_axis

(

int

axis

)

Removes a specified axis.

Parameters

axis	Specified axis

Exceptions

txeo::TensorShapeError

Example Usage:

#include <iostream>
#include "txeo/TensorShape.h"
 
int main() {
    txeo::TensorShape shape({2, 3, 4}); // Initial shape (2,3,4)
    shape.remove_axis(1); // Removes the axis at index 1, resulting in shape (2,4)
 
    std::cout << "Updated TensorShape: " << shape << std::endl;
    return 0;
}

set_dim()

void txeo::TensorShape::set_dim	(	int	axis,
		size_t	dim
	)

Sets a dimension in a specified axis.

Parameters

axis	Axis whose dimension will be changed
dim	New dimension

Exceptions

txeo::TensorShapeError

Example Usage:

#include <iostream>
#include "txeo/TensorShape.h"
 
int main() {
    txeo::TensorShape shape({2, 3, 4}); // Initial shape (2,3,4)
    shape.set_dim(1, 5); // Changes the second axis size from 3 to 5
 
    std::cout << "Updated TensorShape: " << shape << std::endl;
    return 0;
}

size()

int txeo::TensorShape::size ( ) const

inlinenoexcept

Synonym for txeo::TensorShape::number_of_axes()

Returns

int Size of this shape

Definition at line 138 of file TensorShape.h.

{ return this->number_of_axes(); };

Here is the call graph for this function:

stride()

const std::vector< size_t > & txeo::TensorShape::stride

(

)

const

Returns the stride of each dimension in the tensor.

The stride represents the step size needed to move along each dimension of the tensor in memory layout. It is useful for operations requiring efficient indexing.

Returns

The stride for each dimension.

Example Usage:

#include <iostream>
#include "txeo/Tensor.h"
 
int main() {
    txeo::Tensor<int> tensor({3, 4, 5}); // Create a 3x4x5 tensor
    const std::vector<size_t>& strides = tensor.stride();
 
    std::cout << "Tensor strides: ";
    for (size_t s : strides) {
        std::cout << s << " ";
    }
    std::cout << std::endl;
    return 0;
}

Friends And Related Symbol Documentation

operator<<

std::ostream & operator<< ( std::ostream &  os, const TensorShape &  shape  )

friend

Predictor

template<typename T >

friend class Predictor

friend

Definition at line 314 of file TensorShape.h.

Tensor

template<typename T >

friend class Tensor

friend

Definition at line 311 of file TensorShape.h.

TensorAgg

template<typename T >

friend class TensorAgg

friend

Definition at line 317 of file TensorShape.h.

txeo::detail::TensorHelper

friend class txeo::detail::TensorHelper

friend

Definition at line 319 of file TensorShape.h.

---

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

• include/txeo/TensorShape.h