Functions for tensor operations in the C3E library. More...

#include <c3e/commons.h>

Include dependency graph for tensor.h:

This graph shows which files directly or indirectly include this file:

Functions
c3e_tensor *	c3e_tensor_init (size_t dsize, uint32_t dims, c3e_matrix *matrices, c3e_vector data)
	Initializes a tensor with given data and dimensions.

void	c3e_tensor_free (c3e_tensor *tensor)
	Frees the resources associated with a tensor.

c3e_tensor *	c3e_tensor_add (c3e_tensor left, c3e_tensor right)
	Adds two tensors element-wise.

c3e_tensor *	c3e_tensor_sub (c3e_tensor left, c3e_tensor right)
	Subtracts one tensor from another element-wise.

c3e_tensor *	c3e_tensor_mul (c3e_tensor left, c3e_tensor right)
	Multiplies two tensors element-wise.

c3e_tensor *	c3e_tensor_div (c3e_tensor left, c3e_tensor right)
	Divides one tensor by another element-wise.

c3e_tensor *	c3e_tensor_scale (c3e_tensor *tensor, int x)
	Scales the elements of a tensor by a scalar value.

c3e_tensor *	c3e_tensor_exp (c3e_tensor *tensor)
	Applies the exponential function to each element of a tensor.

c3e_tensor *	c3e_tensor_normalize (c3e_tensor *tensor)
	Normalizes the elements of a tensor.

c3e_tensor *	c3e_tensor_copy (c3e_tensor *tensor)
	Creates a copy of a tensor.

c3e_tensor *	c3e_tensor_zeros (size_t dsize, uint32_t dims, int rows, int cols)
	Creates a tensor filled with zeros.

c3e_tensor *	c3e_tensor_ones (size_t dsize, uint32_t dims, int rows, int cols)
	Creates a tensor filled with ones.

c3e_tensor *	c3e_tensor_fill (size_t dsize, uint32_t dims, int rows, int cols, c3e_number value)
	Creates a tensor filled with a specific value.

c3e_tensor *	c3e_tensor_random (size_t dsize, uint32_t dims, int rows, int cols, int seed)
	Creates a tensor with random values within a range.

c3e_tensor *	c3e_tensor_random_bound (size_t dsize, uint32_t dims, int rows, int cols, int seed, c3e_number min, c3e_number max)
	Creates a tensor with random values within a specified range.

bool	c3e_tensor_equals (c3e_tensor tensor, c3e_tensor subject)
	Checks if two tensors are equal.

bool	c3e_tensor_all_close (c3e_tensor tensor, c3e_tensor subject)
	Checks if two tensors are element-wise close within a tolerance.

Detailed Description

Author: Nathanne Isip

This file provides functions for creating, manipulating, and performing operations on tensors. Tensors are multi-dimensional arrays of numerical values, generalizing vectors and matrices. The provided functions include tensor initialization, arithmetic operations, and utilities.

Function Documentation

◆ c3e_tensor_add()

c3e_tensor * c3e_tensor_add	(	c3e_tensor *	left,
		c3e_tensor *	right
	)

This function performs element-wise addition of two tensors with the same dimensions.

Parameters

left	Pointer to the first tensor.
right	Pointer to the second tensor.

Returns: A pointer to a new c3e_tensor structure containing the result of the addition.

◆ c3e_tensor_all_close()

bool c3e_tensor_all_close	(	c3e_tensor *	tensor,
		c3e_tensor *	subject
	)

This function compares two tensors to determine if they are element-wise close within a specified tolerance.

Parameters

tensor	Pointer to the first tensor.
subject	Pointer to the second tensor.

Returns: true if the tensors are element-wise close; false otherwise.

◆ c3e_tensor_copy()

c3e_tensor * c3e_tensor_copy ( c3e_tensor * tensor )

This function duplicates the given tensor, creating a new tensor with the same data and dimensions.

Parameters

tensor Pointer to the tensor to be copied.

Returns: A pointer to a new c3e_tensor structure that is a copy of the original tensor.

◆ c3e_tensor_div()

c3e_tensor * c3e_tensor_div	(	c3e_tensor *	left,
		c3e_tensor *	right
	)

This function performs element-wise division of two tensors with the same dimensions.

Parameters

left	Pointer to the tensor to be divided.
right	Pointer to the tensor to divide by.

Returns: A pointer to a new c3e_tensor structure containing the result of the division.

◆ c3e_tensor_equals()

bool c3e_tensor_equals	(	c3e_tensor *	tensor,
		c3e_tensor *	subject
	)

This function compares two tensors to determine if they are element-wise equal.

Parameters

tensor	Pointer to the first tensor.
subject	Pointer to the second tensor.

Returns: true if the tensors are equal; false otherwise.

◆ c3e_tensor_exp()

c3e_tensor * c3e_tensor_exp ( c3e_tensor * tensor )

This function computes the exponential of every element in the tensor.

Parameters

tensor Pointer to the tensor to be transformed.

Returns: A pointer to a new c3e_tensor structure containing the exponential-transformed tensor.

◆ c3e_tensor_fill()

c3e_tensor * c3e_tensor_fill	(	size_t	dsize,
		uint32_t	dims,
		int	rows,
		int	cols,
		c3e_number	value
	)

This function creates a tensor of specified dimensions and size, initialized with a given value.

Parameters

dsize	The size of the data vector.
dims	The number of dimensions in the tensor.
rows	The number of rows in each matrix.
cols	The number of columns in each matrix.
value	The value to fill the tensor with.

Returns: A pointer to the newly created c3e_tensor structure filled with the specified value.

◆ c3e_tensor_free()

void c3e_tensor_free ( c3e_tensor * tensor )

This function deallocates memory used by the c3e_tensor structure, including memory for the matrices and vector data.

Parameters

tensor Pointer to the c3e_tensor structure to be freed.

◆ c3e_tensor_init()

c3e_tensor * c3e_tensor_init	(	size_t	dsize,
		uint32_t	dims,
		c3e_matrix **	matrices,
		c3e_vector *	data
	)

This function creates a new tensor with the specified number of dimensions and size, initializing it with provided matrices and vector data.

Parameters

dsize	The size of the data vector.
dims	The number of dimensions in the tensor.
matrices	Array of matrices representing the tensor data.
data	Pointer to a vector containing the tensor data.

Returns: A pointer to the newly created c3e_tensor structure.

◆ c3e_tensor_mul()

c3e_tensor * c3e_tensor_mul	(	c3e_tensor *	left,
		c3e_tensor *	right
	)

This function performs element-wise multiplication of two tensors with the same dimensions.

Parameters

left	Pointer to the first tensor.
right	Pointer to the second tensor.

Returns: A pointer to a new c3e_tensor structure containing the result of the multiplication.

◆ c3e_tensor_normalize()

c3e_tensor * c3e_tensor_normalize ( c3e_tensor * tensor )

This function normalizes the tensor elements, typically scaling them to a specific range or unit length.

Parameters

tensor Pointer to the tensor to be normalized.

Returns: A pointer to a new c3e_tensor structure containing the normalized tensor.

◆ c3e_tensor_ones()

c3e_tensor * c3e_tensor_ones	(	size_t	dsize,
		uint32_t	dims,
		int	rows,
		int	cols
	)

This function creates a tensor of specified dimensions and size, initialized with ones.

Parameters

dsize	The size of the data vector.
dims	The number of dimensions in the tensor.
rows	The number of rows in each matrix.
cols	The number of columns in each matrix.

Returns: A pointer to the newly created c3e_tensor structure filled with ones.

◆ c3e_tensor_random()

c3e_tensor * c3e_tensor_random	(	size_t	dsize,
		uint32_t	dims,
		int	rows,
		int	cols,
		int	seed
	)

This function generates a tensor with random values within a specified range, using a given seed for randomness.

Parameters

dsize	The size of the data vector.
dims	The number of dimensions in the tensor.
rows	The number of rows in each matrix.
cols	The number of columns in each matrix.
seed	The seed value for the random number generator.
min	The minimum value of the random range.
max	The maximum value of the random range.

Returns: A pointer to the newly created c3e_tensor structure filled with random values.

◆ c3e_tensor_random_bound()

c3e_tensor * c3e_tensor_random_bound	(	size_t	dsize,
		uint32_t	dims,
		int	rows,
		int	cols,
		int	seed,
		c3e_number	min,
		c3e_number	max
	)

This function generates a tensor with random values within a specified range, using a given seed for randomness.

Parameters

dsize	The size of the data vector.
dims	The number of dimensions in the tensor.
rows	The number of rows in each matrix.
cols	The number of columns in each matrix.
seed	The seed value for the random number generator.
min	The minimum value of the random range.
max	The maximum value of the random range.

Returns: A pointer to the newly created c3e_tensor structure filled with random values within the specified range.

◆ c3e_tensor_scale()

c3e_tensor * c3e_tensor_scale	(	c3e_tensor *	tensor,
		int	x
	)

This function multiplies every element of the tensor by a specified scalar value.

Parameters

tensor	Pointer to the tensor to be scaled.
x	The scalar value by which to scale the tensor.

Returns: A pointer to a new c3e_tensor structure containing the scaled tensor.

◆ c3e_tensor_sub()

c3e_tensor * c3e_tensor_sub	(	c3e_tensor *	left,
		c3e_tensor *	right
	)

This function performs element-wise subtraction of two tensors with the same dimensions.

Parameters

left	Pointer to the tensor to be subtracted from.
right	Pointer to the tensor to subtract.

Returns: A pointer to a new c3e_tensor structure containing the result of the subtraction.

◆ c3e_tensor_zeros()

c3e_tensor * c3e_tensor_zeros	(	size_t	dsize,
		uint32_t	dims,
		int	rows,
		int	cols
	)

This function creates a tensor of specified dimensions and size, initialized with zeros.

Parameters

dsize	The size of the data vector.
dims	The number of dimensions in the tensor.
rows	The number of rows in each matrix.
cols	The number of columns in each matrix.

Returns: A pointer to the newly created c3e_tensor structure filled with zeros.