TensorTT

# Import statement

from hottbox.core import TensorTT

class TensorTT(core_values, mode_names=None)

Representation of a tensor in the Tensor Train form (TT).

Parameters

core_valueslist[np.ndarray]

List of cores for the Tensor Train representation of a tensor.

mode_nameslist[str]

List of names for the physical modes

Attributes

_core_values	(list[np.ndarray]) Placeholder for a list of cores for the Tensor Train representation of a tensor.
_modes	(list[Mode]) Description of the physical modes

Raises

TensorTopologyError

If there is inconsistency in shapes of core values

Examples

1. Create tensor train representation of a tensor with default meta information

>>> import numpy as np
>>> from hottbox.core import TensorTT
>>> I, J, K = 5, 6, 7          # shape of the tensor in full form
>>> R_1, R_2 = 3, 4            # tt-rank
>>> core_0 = np.ones((I, R_1))
>>> core_1 = np.ones((R1, J, R2))
>>> core_2 = np.ones((R2, K))
>>> core_values = [core_0, core_1, core_2]
>>> tensor_tt = TensorTT(core_values)
>>> print(tensor_tt)
    Tensor train representation of a tensor with tt-rank=(3, 4).
    Shape of this representation in the full format is (5, 6, 7).
    Physical modes of its cores represent properties: ['mode-0', 'mode-1', 'mode-2']

2. Create tensor train representation of a tensor with custom meta information

>>> import numpy as np
>>> from hottbox.core import TensorTT
>>> I, J, K = 5, 6, 7  # shape of the tensor in full form
>>> R_1, R_2 = 3, 4    # tt-rank
>>> core_0 = np.ones((I, R_1))
>>> core_1 = np.ones((R_1, J, R_2))
>>> core_2 = np.ones((R_2, K))
>>> core_values = [core_0, core_1, core_2]
>>> mode_names = ["Year", "Month", "Day"]
>>> tensor_tt = TensorTT(core_values, mode_names)
>>> print(tensor_tt)
    Tensor train representation of a tensor with tt-rank=(3, 4).
    Shape of this representation in the full format is (5, 6, 7).
    Physical modes of its cores represent properties: ['Year', 'Month', 'Day']

Methods

copy(self)	Produces a copy of itself as a new object
copy_modes(self, tensor)	Copy modes meta from tensor
core(self, i)	Specific core of the TensorTT representation
reconstruct(self[, keep_meta])	Converts the TT representation of a tensor into a full tensor
reset_mode_index(self[, mode])	Drop index for the specified mode number
reset_mode_name(self[, mode])	Set default name for the specified mode number
set_mode_index(self, mode_index)	Set index for specified mode
set_mode_names(self, mode_names)	Rename modes of a tensor representation

copy(self)

Produces a copy of itself as a new object

Returns

new_objectTensorTT

copy_modes(self, tensor)

Copy modes meta from tensor

Parameters

tensor{Tensor, TensorCPD, TensorTKD, TensorTT}

Returns

selfTensorTT

Notes

Most of the time this method should only be used by the decomposition algorithms

core(self, i)

Specific core of the TensorTT representation

Parameters

iint

Should not exceed the order of TensorTT.order representation

Returns

core_tensorTensor

property cores

All cores of the TensorTT representation

Returns

core_listlist[Tensor]

property ft_shape

Shape of a TensorTT in the full format

Returns

full_shapetuple

property mode_names

Description of the physical modes for a TensorTT

Returns

list[str]

property modes

Meta data for the factor matrices

Returns

list[Mode]

property order

Order of a tensor represented through the TT

Returns

orderint

property rank

Rank of the TT representation of a tensor

Returns

ranktuple

Notes

Most often referred to as the TT rank

reconstruct(self, keep_meta=0)

Converts the TT representation of a tensor into a full tensor

Parameters

keep_metaint

Keep meta information about modes of the given tensor. 0 - the output will have default values for the meta data 1 - keep only mode names 2 - keep mode names and indices

Returns

tensorTensor

reset_mode_index(self, mode=None)

Drop index for the specified mode number

Parameters

modeint

Mode number which index to be dropped By default resets all indices

Returns

selfTensorTT

reset_mode_name(self, mode=None)

Set default name for the specified mode number

Parameters

modeint

Mode number which name to be set to default value By default resets names of all modes

Returns

selfTensorTT

set_mode_index(self, mode_index)

Set index for specified mode

Parameters

mode_indexdict

New indices for the factor matrices in form of a dictionary. Key defines the mode whose index to be changed. Value contains a list of new indices for this mode.

Returns

selfTensorTT

set_mode_names(self, mode_names)

Rename modes of a tensor representation

Parameters

mode_namesdict

New names for the tensor modes in form of a dictionary The name of the mode defined by the Key of the dict will be renamed to the corresponding Value

Returns

selfTensorTT