A single-batched, logically 2D LabeledTensor. More...

Detailed Description

A single-batched, logically 2D LabeledTensor.

#include <LabeledMatrix.h>

Inheritance diagram for LabeledMatrix:

Public Member Functions
void	accumulate (const LabeledMatrix &other, bool recursive=true)

void	fill (const LabeledMatrix &other, bool recursive=true)

LabeledMatrix	chain (const LabeledMatrix &other) const
	Chain rule product of two derivatives.

LabeledMatrix	inverse () const
	Invert a LabeledMatrix for use in an implicit function derivative.

Public Member Functions inherited from LabeledTensor< LabeledMatrix, 2 >
	LabeledTensor ()=default
	Default constructor.

	LabeledTensor (const torch::Tensor &tensor, TorchSize batch_dim, const std::vector< const LabeledAxis * > &axes)
	Construct from a Tensor with batch dim and vector of `LabeledAxis`

	LabeledTensor (const BatchTensor &tensor, const std::vector< const LabeledAxis * > &axes)
	Construct from a BatchTensor with vector of `LabeledAxis`

	LabeledTensor (const LabeledMatrix &other)
	Copy constructor.

void	operator= (const LabeledMatrix &other)
	Assignment operator.

	operator BatchTensor () const
	A potentially dangerous implicit conversion.

	operator torch::Tensor () const
	A potentially dangerous implicit conversion.

LabeledMatrix	clone (torch::MemoryFormat memory_format=torch::MemoryFormat::Contiguous) const
	Clone this LabeledTensor.

void	copy_ (const T &other)
	Copy the value from another tensor.

LabeledMatrix	detach () const
	Return a copy without gradient graphs.

void	detach_ ()
	Detach from gradient graphs.

void	zero_ ()
	Zero out this tensor.

torch::TensorOptions	options () const
	Get the tensor options.

TorchSize	batch_dim () const
	Return the number of batch dimensions.

TorchSize	base_dim () const
	Return the number of base dimensions.

TorchShapeRef	batch_sizes () const
	Return the batch size.

TorchShapeRef	base_sizes () const
	Return the base size.

const std::vector< const LabeledAxis * > &	axes () const
	Get all the labeled axes.

const LabeledAxis &	axis (TorchSize i=0) const
	Get a specific labeled axis.

TorchSlice	slice_indices (S &&... names) const
	How to slice the tensor given the names on each axis.

TorchShapeRef	storage_size () const
	The shape of the entire LabeledTensor.

TorchShape	storage_size (S &&... names) const
	The shape of a sub-block specified by the names on each dimension.

BatchTensor	operator() (S &&... names) const

LabeledMatrix	slice (TorchSize i, const std::string &name) const
	Slice the tensor on the given dimension by a single variable or sub-axis.

LabeledMatrix	block (S &&... names) const
	Get the sub-block labeled by the given sub-axis names.

LabeledMatrix	batch_index (TorchSlice indices) const
	Get a batch.

void	batch_index_put (TorchSlice indices, const torch::Tensor &other)
	Set a index sliced on the batch dimensions to a value.

BatchTensor	base_index (TorchSlice indices) const
	Return an index sliced on the batch dimensions.

void	base_index_put (TorchSlice indices, const torch::Tensor &other)
	Set a index sliced on the batch dimensions to a value.

variable_type< T >::type	get (S &&... names) const
	Get and interpret the view as an object.

variable_type< T >::type	get_list (S &&... names) const
	Get and interpret the view as a list of objects.

void	set (const BatchTensorBase< T > &value, S &&... names)
	Set and interpret the input as an object.

void	set_list (const BatchTensorBase< T > &value, S &&... names)
	Set and interpret the input as a list of objects.

LabeledMatrix	operator- () const
	Negation.

LabeledMatrix	to (const torch::TensorOptions &options) const
	Change tensor options.

const BatchTensor &	tensor () const

BatchTensor &	tensor ()

Static Public Member Functions
static LabeledMatrix	identity (TorchShapeRef batch_size, const LabeledAxis &axis, const torch::TensorOptions &options=default_tensor_options())
	Create a labeled identity tensor.

Static Public Member Functions inherited from LabeledTensor< LabeledMatrix, 2 >
static LabeledMatrix	empty (TorchShapeRef batch_shape, const std::vector< const LabeledAxis * > &axes, const torch::TensorOptions &options=default_tensor_options())
	Setup new empty storage.

static LabeledMatrix	empty_like (const LabeledMatrix &other)
	Setup new empty storage like another LabeledTensor.

static LabeledMatrix	zeros (TorchShapeRef batch_shape, const std::vector< const LabeledAxis * > &axes, const torch::TensorOptions &options=default_tensor_options())
	Setup new storage with zeros.

static LabeledMatrix	zeros_like (const LabeledMatrix &other)
	Setup new storage with zeros like another LabeledTensor.

Additional Inherited Members
Protected Attributes inherited from LabeledTensor< LabeledMatrix, 2 >
BatchTensor	_tensor
	The tensor.

std::vector< const LabeledAxis * >	_axes
	The labeled axes of this tensor.

Member Function Documentation

◆ accumulate()

void accumulate	(	const LabeledMatrix &	other,
		bool	recursive = true )

Add another matrix into this matrix. The item set of the other matrix must be a subset of this matrix's item set.

◆ chain()

LabeledMatrix chain ( const LabeledMatrix & other ) const

Chain rule product of two derivatives.

◆ fill()

void fill	(	const LabeledMatrix &	other,
		bool	recursive = true )

Fill another matrix into this matrix. The item set of the other matrix must be a subset of this matrix's item set.

◆ identity()

LabeledMatrix identity	(	TorchShapeRef	batch_size,
		const LabeledAxis &	axis,
		const torch::TensorOptions &	options = default_tensor_options() )

static

Create a labeled identity tensor.

◆ inverse()

LabeledMatrix inverse ( ) const

Invert a LabeledMatrix for use in an implicit function derivative.