tf.TypeSpec

Specifies a TensorFlow value type.

Inherits From: TraceType

A tf.TypeSpec provides metadata describing an object accepted or returned by TensorFlow APIs. Concrete subclasses, such as tf.TensorSpec and tf.RaggedTensorSpec, are used to describe different value types.

For example, tf.function's input_signature argument accepts a list (or nested structure) of TypeSpecs.

Creating new subclasses of TypeSpec (outside of TensorFlow core) is not currently supported. In particular, we may make breaking changes to the private methods and properties defined by this base class.

Example:

spec = tf.RaggedTensorSpec(shape=[None, None], dtype=tf.int32)
@tf.function(input_signature=[spec])
def double(x):
  return x * 2
print(double(tf.ragged.constant([[1, 2], [3]])))
<tf.RaggedTensor [[2, 4], [6]]>

value_type The Python type for values that are compatible with this TypeSpec.

In particular, all values that are compatible with this TypeSpec must be an instance of this type.

Methods

experimental_as_proto

View source

experimental_as_proto() -> struct_pb2.TypeSpecProto

Returns a proto representation of the TypeSpec instance.

Do NOT override for custom non-TF types.

experimental_from_proto

View source

@classmethod
experimental_from_proto(
    proto: struct_pb2.TypeSpecProto
) -> 'TypeSpec'

Returns a TypeSpec instance based on the serialized proto.

Do NOT override for custom non-TF types.

Args
proto Proto generated using 'experimental_as_proto'.

experimental_type_proto

View source

@classmethod
experimental_type_proto() -> Type[struct_pb2.TypeSpecProto]

Returns the type of proto associated with TypeSpec serialization.

Do NOT override for custom non-TF types.

is_compatible_with

View source

is_compatible_with(
    spec_or_value
)

Returns true if spec_or_value is compatible with this TypeSpec.

Prefer using "is_subtype_of" and "most_specific_common_supertype" wherever possible.

Args
spec_or_value A TypeSpec or TypeSpec associated value to compare against.

is_subtype_of

View source

is_subtype_of(
    other: tf.types.experimental.TraceType
) -> bool

Returns True if self is a subtype of other.

Implements the tf.types.experimental.func.TraceType interface.

If not overridden by a subclass, the default behavior is to assume the TypeSpec is covariant upon attributes that implement TraceType and invariant upon rest of the attributes as well as the structure and type of the TypeSpec.

Args
other A TraceType object.

most_specific_common_supertype

View source

most_specific_common_supertype(
    others: Sequence[tf.types.experimental.TraceType]
) -> Optional['TypeSpec']

Returns the most specific supertype TypeSpec of self and others.

Implements the tf.types.experimental.func.TraceType interface.

If not overridden by a subclass, the default behavior is to assume the TypeSpec is covariant upon attributes that implement TraceType and invariant upon rest of the attributes as well as the structure and type of the TypeSpec.

Args
others A sequence of TraceTypes.

most_specific_compatible_type

View source

most_specific_compatible_type(
    other: 'TypeSpec'
) -> 'TypeSpec'

Returns the most specific TypeSpec compatible with self and other. (deprecated)

Deprecated. Please use most_specific_common_supertype instead. Do not override this function.

Args
other A TypeSpec.

Raises
ValueError If there is no TypeSpec that is compatible with both self and other.

__eq__

View source

__eq__(
    other
) -> bool

Return self==value.

__ne__

View source

__ne__(
    other
) -> bool

Return self!=value.