FullBatchTensor Class Reference

Create a full BatchTensor from the input file. More...

Detailed Description

Create a full BatchTensor from the input file.

#include <FullBatchTensor.h>

Inheritance diagram for FullBatchTensor:

Public Member Functions
	FullBatchTensor (const OptionSet &options)
	Construct a new FullBatchTensor object.
Public Member Functions inherited from BatchTensorBase< BatchTensor >
	BatchTensorBase ()=default
	Default constructor.
	BatchTensorBase (const torch::Tensor &tensor, TorchSize batch_dim)
	Construct from another torch::Tensor.
	BatchTensorBase (const BatchTensor &tensor)
	Copy constructor.
	BatchTensorBase (Real)=delete
bool	batched () const
	Whether the tensor is batched.
TorchSize	batch_dim () const
	Return the number of batch dimensions.
TorchSize &	batch_dim ()
	Return a writable reference to the batch dimension.
TorchSize	base_dim () const
	Return the number of base dimensions.
TorchShapeRef	batch_sizes () const
	Return the batch size.
TorchSize	batch_size (TorchSize index) const
	Return the length of some batch axis.
TorchShapeRef	base_sizes () const
	Return the base size.
TorchSize	base_size (TorchSize index) const
	Return the length of some base axis.
TorchSize	base_storage () const
	Return the flattened storage needed just for the base indices.
BatchTensor	batch_index (TorchSlice indices) const
	Get a batch.
BatchTensor	base_index (const TorchSlice &indices) const
	Return an index sliced on the base dimensions.
void	batch_index_put (TorchSlice indices, const torch::Tensor &other)
	Set a index sliced on the batch dimensions to a value.
void	base_index_put (const TorchSlice &indices, const torch::Tensor &other)
	Set a index sliced on the base dimensions to a value.
BatchTensor	batch_expand (TorchShapeRef batch_size) const
	Return a new view of the tensor with values broadcast along the batch dimensions.
BatchTensor	base_expand (TorchShapeRef base_size) const
	Return a new view of the tensor with values broadcast along the base dimensions.
BatchTensor	batch_expand_as (const Derived2 &other) const
	Expand the batch to have the same shape as another tensor.
Derived2	base_expand_as (const Derived2 &other) const
	Expand the base to have the same shape as another tensor.
BatchTensor	batch_expand_copy (TorchShapeRef batch_size) const
	Return a new tensor with values broadcast along the batch dimensions.
BatchTensor	base_expand_copy (TorchShapeRef base_size) const
	Return a new tensor with values broadcast along the base dimensions.
BatchTensor	batch_reshape (TorchShapeRef batch_shape) const
	Reshape batch dimensions.
BatchTensor	base_reshape (TorchShapeRef base_shape) const
	Reshape base dimensions.
BatchTensor	batch_unsqueeze (TorchSize d) const
	Unsqueeze a batch dimension.
BatchTensor	list_unsqueeze () const
	Unsqueeze on the special list batch dimension.
BatchTensor	base_unsqueeze (TorchSize d) const
	Unsqueeze a base dimension.
BatchTensor	batch_transpose (TorchSize d1, TorchSize d2) const
	Transpose two batch dimensions.
BatchTensor	base_transpose (TorchSize d1, TorchSize d2) const
	Transpose two base dimensions.
BatchTensor	base_movedim (TorchSize d1, TorchSize d2) const
	Move two base dimensions.
BatchTensor	clone (torch::MemoryFormat memory_format=torch::MemoryFormat::Contiguous) const
	Clone (take ownership)
BatchTensor	detach () const
	Discard function graph.
BatchTensor	to (const torch::TensorOptions &options) const
	Send to options.
BatchTensor	operator- () const
	Negation.
BatchTensor	batch_sum (TorchSize d) const
	Sum on a batch index.
BatchTensor	list_sum () const
	Sum on the list index (TODO: replace with class)
Public Member Functions inherited from UserTensor
	UserTensor (const OptionSet &options)
Public Member Functions inherited from NEML2Object
	NEML2Object (const OptionSet &options)
	Construct a new NEML2Object object.
virtual	~NEML2Object ()=default
const OptionSet &	input_options () const
virtual void	setup ()
	Setup this object.
const std::string &	name () const
	A readonly reference to the object's name.
const std::string &	type () const
	A readonly reference to the object's type.
const std::string &	path () const
	A readonly reference to the object's path.
const std::string &	doc () const
	A readonly reference to the object's docstring.
template<typename T = NEML2Object>
const T *	host () const
	Get a readonly pointer to the host.
template<typename T = NEML2Object>
T *	host ()
	Get a writable pointer to the host.

Static Public Member Functions
static OptionSet	expected_options ()
Static Public Member Functions inherited from BatchTensor
static BatchTensor	empty (const TorchShapeRef &base_shape, const torch::TensorOptions &options=default_tensor_options())
	Unbatched empty tensor given base shape.
static BatchTensor	empty (const TorchShapeRef &batch_shape, const TorchShapeRef &base_shape, const torch::TensorOptions &options=default_tensor_options())
	Empty tensor given batch and base shapes.
static BatchTensor	zeros (const TorchShapeRef &base_shape, const torch::TensorOptions &options=default_tensor_options())
	Unbatched tensor filled with zeros given base shape.
static BatchTensor	zeros (const TorchShapeRef &batch_shape, const TorchShapeRef &base_shape, const torch::TensorOptions &options=default_tensor_options())
	Zero tensor given batch and base shapes.
static BatchTensor	ones (const TorchShapeRef &base_shape, const torch::TensorOptions &options=default_tensor_options())
	Unbatched tensor filled with ones given base shape.
static BatchTensor	ones (const TorchShapeRef &batch_shape, const TorchShapeRef &base_shape, const torch::TensorOptions &options=default_tensor_options())
	Unit tensor given batch and base shapes.
static BatchTensor	full (const TorchShapeRef &base_shape, Real init, const torch::TensorOptions &options=default_tensor_options())
	Unbatched tensor filled with a given value given base shape.
static BatchTensor	full (const TorchShapeRef &batch_shape, const TorchShapeRef &base_shape, Real init, const torch::TensorOptions &options=default_tensor_options())
	Full tensor given batch and base shapes.
static BatchTensor	identity (TorchSize n, const torch::TensorOptions &options=default_tensor_options())
	Unbatched identity tensor.
static BatchTensor	identity (const TorchShapeRef &batch_shape, TorchSize n, const torch::TensorOptions &options=default_tensor_options())
	Identity tensor given batch shape and base length.
Static Public Member Functions inherited from BatchTensorBase< BatchTensor >
static BatchTensor	empty_like (const BatchTensor &other)
	Empty tensor like another, i.e. same batch and base shapes, same tensor options, etc.
static BatchTensor	zeros_like (const BatchTensor &other)
	Zero tensor like another, i.e. same batch and base shapes, same tensor options, etc.
static BatchTensor	ones_like (const BatchTensor &other)
	Unit tensor like another, i.e. same batch and base shapes, same tensor options, etc.
static BatchTensor	full_like (const BatchTensor &other, Real init)
static BatchTensor	linspace (const BatchTensor &start, const BatchTensor &end, TorchSize nstep, TorchSize dim=0, TorchSize batch_dim=-1)
	Create a new tensor by adding a new batch dimension with linear spacing between start and end.
static BatchTensor	logspace (const BatchTensor &start, const BatchTensor &end, TorchSize nstep, TorchSize dim=0, TorchSize batch_dim=-1, Real base=10)
	log-space equivalent of the linspace named constructor
Static Public Member Functions inherited from UserTensor
static OptionSet	expected_options ()
Static Public Member Functions inherited from NEML2Object
static OptionSet	expected_options ()

Constructor & Destructor Documentation

FullBatchTensor()

FullBatchTensor

(

const OptionSet &

options

)

Construct a new FullBatchTensor object.

Parameters

options

The options extracted from the input file.

Member Function Documentation

expected_options()

OptionSet expected_options ( )

static