concepts.dsl.tensor_state.MultidimensionalArrayInterface#

class MultidimensionalArrayInterface[source]#

Bases: object

A multi-dimensional array inferface. At a high-level, this can be interpreted as a dictionary that maps feature names (keys) to multi-diemsntional tensors (value).

Methods

`clone`()	Clone the multidimensional array interface.
`get_feature`(name)	Get the feature tensor with the given name.
`items`()	Get the feature name-tensor pairs.
`keys`()	Get the feature names.
`set_feature`(name, feature)	Set the feature tensor with the given name.
`update_feature`(other_tensor_dict)	Update the feature tensors with the given tensor dict.
`values`()	Get the feature tensors.

__getitem__(name)[source]#

Get the feature tensor with the given name.

Parameters:: name (str)
Return type:: TensorValue

__init__(all_feature_names=tuple())[source]#

Parameters:: all_feature_names (Iterable[str])

__new__(**kwargs)#

clone()[source]#

Clone the multidimensional array interface.

Return type:: MultidimensionalArrayInterface

get_feature(name)[source]#

Get the feature tensor with the given name.

Parameters:: name (str)
Return type:: TensorValue

items()[source]#

Get the feature name-tensor pairs.

Return type:: Iterable[Tuple[str, TensorValue]]

keys()[source]#

Get the feature names.

Return type:: Iterable[str]

set_feature(name, feature)[source]#

Set the feature tensor with the given name.

Parameters:

name (str)
feature (TensorValue)

update_feature(other_tensor_dict)[source]#

Update the feature tensors with the given tensor dict.

Parameters:: other_tensor_dict (Mapping[str, TensorValue])

values()[source]#

Get the feature tensors.

Return type:: Iterable[TensorValue]