Manipulation Class Reference

Compute essential tensor operations. More...

Inheritance diagram for Manipulation:

Public Member Functions
void	Collapse (int collapse_mode, char *in_tensor_path, char *out_tensor_path, const GPU_MODE _gpu_mode=ANY)
	Collapsing a mode of a tensor. More...
void	ConvertSign (char *in_tensor_path, char *out_tensor_path, const GPU_MODE _gpu_mode=ANY)
	Converting signs of a tensor. More...
void	ConvertToBinaryTensor (char *in_tensor_path, char *out_tensor_path, const GPU_MODE _gpu_mode=ANY)
	Converting a tensor into a binary tensor. More...
void	Find (int *index, char *in_tensor_path, char *out_tensor_path, const GPU_MODE _gpu_mode=ANY)
	Finding elements satisfying certain conditions or index in a tensor. More...
void	Matricize (int mode, char *in_tensor_path, char *out_tensor_path, const GPU_MODE _gpu_mode=ANY)
	Matricization of a tensor to certain mode. More...
void	Fold (char *in_tensor_path, int order, int mode, int *dim, char *out_tensor_path, const GPU_MODE _gpu_mode=ANY)
	Converting a matrix into a tensor. More...
void	Norm (char type, char *in_tensor_path, const GPU_MODE _gpu_mode=ANY)
	Computing L1-norm or L-F norm of a tensor. More...
void	PermutateModes (int *modes, char *in_tensor_path, char *out_tensor_path, const GPU_MODE _gpu_mode=ANY)
	Permuting the order of modes of a tensor (transpose in a matrix) More...
void	ScalarOP (char *in_tensor_path, char op, const float val, char *out_tensor_path, const GPU_MODE _gpu_mode)
	Execution of scalar operations (+, -, *, /) More...
void	Scale (char *in_tensor_path, const int new_mode, char *out_tensor_path, const GPU_MODE _gpu_mode)
	Expansion of a tensor to given a new mode. More...

Detailed Description

Compute essential tensor operations.

Member Function Documentation

Collapse()

void Manipulation::Collapse	(	int	collapse_mode,
		char *	in_tensor_path,
		char *	out_tensor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Collapsing a mode of a tensor.

Parameters

collapse_mode	mode of a tensor to collapse
in_tensor_path	input tensor path
out_tensor_path	output tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

ConvertSign()

void Manipulation::ConvertSign	(	char *	in_tensor_path,
		char *	out_tensor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Converting signs of a tensor.

Parameters

in_tensor_path	input tensor path
out_tensor_path	output tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

ConvertToBinaryTensor()

void Manipulation::ConvertToBinaryTensor	(	char *	in_tensor_path,
		char *	out_tensor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Converting a tensor into a binary tensor.

Parameters

in_tensor_path	input tensor path
out_tensor_path	output tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

Find()

void Manipulation::Find	(	int *	index,
		char *	in_tensor_path,
		char *	out_tensor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Finding elements satisfying certain conditions or index in a tensor.

Parameters

index	index in the tensor
in_tensor_path	input tensor path
out_tensor_path	output tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

Fold()

void Manipulation::Fold	(	char *	in_csr_path,
		int	order,
		int	mode,
		int *	dim,
		char *	out_tensor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Converting a matrix into a tensor.

Parameters

in_csr_path	input tensor path
order	the order of a tensor
mode	the mode of a tensor to fold
dim	the dimension of a tensor
out_tensor_path	output tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

Matricize()

void Manipulation::Matricize	(	int	mode,
		char *	in_tensor_path,
		char *	out_tensor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Matricization of a tensor to certain mode.

Parameters

mode	the mode of tensor
in_tensor_path	input tensor path
out_tensor_path	output tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

Norm()

void Manipulation::Norm	(	char	type,
		char *	in_tensor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Computing L1-norm or L-F norm of a tensor.

Parameters

type	type of normalization
in_tensor_path	input tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

PermutateModes()

void Manipulation::PermutateModes	(	int *	modes,
		char *	in_tensor_path,
		char *	out_tensor_path,
		const GPU_MODE	_gpu_mode = ANY
	)

Permuting the order of modes of a tensor (transpose in a matrix)

Parameters

modes	list of modes to permutate
in_tensor_path	input tensor path
out_tensor_path	output tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

ScalarOP()

void Manipulation::ScalarOP	(	char *	in_tensor_path,
		char	op,
		const float	val,
		char *	out_tensor_path,
		const GPU_MODE	_gpu_mode
	)

Execution of scalar operations (+, -, *, /)

Parameters

in_tensor_path	input tensor path
op	operator (supporting: +, -, *, /)
val	operation value (e.g.: val is 0.3 in 'x + 0.3')
out_tensor_path	output tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None

Scale()

void Manipulation::Scale	(	char *	in_tensor_path,
		const int	new_mode,
		char *	out_tensor_path,
		const GPU_MODE	_gpu_mode
	)

Expansion of a tensor to given a new mode.

Parameters

in_tensor_path	input tensor path
new_mode	new mode
out_tensor_path	output tensor path
_gpu_mode	gpu mode for this computation (default: gpu_mode of a BIGtensor instance)

Returns

None

Exceptions

None