tf.saved_model - LostTech.TensorFlow Documentation

object build_signature_def(IDictionary<object, object> inputs, IDictionary<object, object> outputs, string method_name)

Utility function to build a SignatureDef protocol buffer.

Parameters

IDictionary<object, object> inputs: Inputs of the SignatureDef defined as a proto map of string to tensor info.
IDictionary<object, object> outputs: Outputs of the SignatureDef defined as a proto map of string to tensor info.
string method_name: Method name of the SignatureDef as a string.

Returns

object: A SignatureDef protocol buffer constructed based on the supplied arguments.

object build_signature_def_dyn(object inputs, object outputs, object method_name)

Utility function to build a SignatureDef protocol buffer.

Parameters

object inputs: Inputs of the SignatureDef defined as a proto map of string to tensor info.
object outputs: Outputs of the SignatureDef defined as a proto map of string to tensor info.
object method_name: Method name of the SignatureDef as a string.

Returns

object: A SignatureDef protocol buffer constructed based on the supplied arguments.

object build_tensor_info(IGraphNodeBase tensor)

Utility function to build TensorInfo proto from a Tensor. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info.

Parameters

IGraphNodeBase tensor: Tensor or SparseTensor whose name, dtype and shape are used to build the TensorInfo. For SparseTensors, the names of the three constituent Tensors are used.

Returns

object: A TensorInfo protocol buffer constructed based on the supplied argument.

object build_tensor_info(string tensor)

Utility function to build TensorInfo proto from a Tensor. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info.

Parameters

string tensor: Tensor or SparseTensor whose name, dtype and shape are used to build the TensorInfo. For SparseTensors, the names of the three constituent Tensors are used.

Returns

object: A TensorInfo protocol buffer constructed based on the supplied argument.

object build_tensor_info(IEnumerable<object> tensor)

Utility function to build TensorInfo proto from a Tensor. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info.

Parameters

IEnumerable<object> tensor: Tensor or SparseTensor whose name, dtype and shape are used to build the TensorInfo. For SparseTensors, the names of the three constituent Tensors are used.

Returns

object: A TensorInfo protocol buffer constructed based on the supplied argument.

object build_tensor_info(RaggedTensor tensor)

Utility function to build TensorInfo proto from a Tensor. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info.

Parameters

RaggedTensor tensor: Tensor or SparseTensor whose name, dtype and shape are used to build the TensorInfo. For SparseTensors, the names of the three constituent Tensors are used.

Returns

object: A TensorInfo protocol buffer constructed based on the supplied argument.

object build_tensor_info(int tensor)

Utility function to build TensorInfo proto from a Tensor. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info.

Parameters

int tensor: Tensor or SparseTensor whose name, dtype and shape are used to build the TensorInfo. For SparseTensors, the names of the three constituent Tensors are used.

Returns

object: A TensorInfo protocol buffer constructed based on the supplied argument.

object build_tensor_info_dyn(object tensor)

Utility function to build TensorInfo proto from a Tensor. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info.

Parameters

object tensor: Tensor or SparseTensor whose name, dtype and shape are used to build the TensorInfo. For SparseTensors, the names of the three constituent Tensors are used.

Returns

object: A TensorInfo protocol buffer constructed based on the supplied argument.

object classification_signature_def(IGraphNodeBase examples, IGraphNodeBase classes, IGraphNodeBase scores)

Creates classification signature from given examples and predictions.

This function produces signatures intended for use with the TensorFlow Serving Classify API (tensorflow_serving/apis/prediction_service.proto), and so constrains the input and output types to those allowed by TensorFlow Serving.

Parameters

IGraphNodeBase examples: A string `Tensor`, expected to accept serialized tf.Examples.
IGraphNodeBase classes: A string `Tensor`. Note that the ClassificationResponse message requires that class labels are strings, not integers or anything else.
IGraphNodeBase scores: a float `Tensor`.

Returns

object: A classification-flavored signature_def.

object classification_signature_def(IGraphNodeBase examples, IGraphNodeBase classes, string scores)

Creates classification signature from given examples and predictions.

This function produces signatures intended for use with the TensorFlow Serving Classify API (tensorflow_serving/apis/prediction_service.proto), and so constrains the input and output types to those allowed by TensorFlow Serving.

Parameters

IGraphNodeBase examples: A string `Tensor`, expected to accept serialized tf.Examples.
IGraphNodeBase classes: A string `Tensor`. Note that the ClassificationResponse message requires that class labels are strings, not integers or anything else.
string scores: a float `Tensor`.

Returns

object: A classification-flavored signature_def.

object classification_signature_def_dyn(object examples, object classes, object scores)

Creates classification signature from given examples and predictions.

This function produces signatures intended for use with the TensorFlow Serving Classify API (tensorflow_serving/apis/prediction_service.proto), and so constrains the input and output types to those allowed by TensorFlow Serving.

Parameters

object examples: A string `Tensor`, expected to accept serialized tf.Examples.
object classes: A string `Tensor`. Note that the ClassificationResponse message requires that class labels are strings, not integers or anything else.
object scores: a float `Tensor`.

Returns

object: A classification-flavored signature_def.

bool contains_saved_model(string export_dir)

Checks whether the provided export directory could contain a SavedModel.

Note that the method does not load any data by itself. If the method returns `false`, the export directory definitely does not contain a SavedModel. If the method returns `true`, the export directory may contain a SavedModel but provides no guarantee that it can be loaded.

Parameters

string export_dir: Absolute string path to possible export location. For example, '/my/foo/model'.

Returns

bool: True if the export directory contains SavedModel files, False otherwise.

SparseTensor get_tensor_from_tensor_info(object tensor_info, Graph graph, string import_scope)

Returns the Tensor or CompositeTensor described by a TensorInfo proto. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.get_tensor_from_tensor_info or tf.compat.v1.saved_model.get_tensor_from_tensor_info.

Parameters

object tensor_info: A TensorInfo proto describing a Tensor or SparseTensor or CompositeTensor.
Graph graph: The tf.Graph in which tensors are looked up. If None, the current default graph is used.
string import_scope: If not None, names in `tensor_info` are prefixed with this string before lookup.

Returns

SparseTensor: The Tensor or SparseTensor or CompositeTensor in `graph` described by `tensor_info`.

object get_tensor_from_tensor_info_dyn(object tensor_info, object graph, object import_scope)

Returns the Tensor or CompositeTensor described by a TensorInfo proto. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.get_tensor_from_tensor_info or tf.compat.v1.saved_model.get_tensor_from_tensor_info.

Parameters

object tensor_info: A TensorInfo proto describing a Tensor or SparseTensor or CompositeTensor.
object graph: The tf.Graph in which tensors are looked up. If None, the current default graph is used.
object import_scope: If not None, names in `tensor_info` are prefixed with this string before lookup.

Returns

object: The Tensor or SparseTensor or CompositeTensor in `graph` described by `tensor_info`.

bool is_valid_signature(object signature_def)

Determine whether a SignatureDef can be served by TensorFlow Serving.

object is_valid_signature_dyn(object signature_def)

Determine whether a SignatureDef can be served by TensorFlow Serving.

AutoTrackable load(Session sess, IEnumerable<string> tags, string export_dir, string import_scope, IDictionary<string, object> saver_kwargs)

Loads the model from a SavedModel as specified by tags. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.loader.load or tf.compat.v1.saved_model.load. There will be a new function for importing SavedModels in Tensorflow 2.0.

Parameters

Session sess: The TensorFlow session to restore the variables.
IEnumerable<string> tags: Set of string tags to identify the required MetaGraphDef. These should correspond to the tags used when saving the variables using the SavedModel `save()` API.
string export_dir: Directory in which the SavedModel protocol buffer and variables to be loaded are located.
string import_scope: Optional `string` -- if specified, prepend this string followed by '/' to all loaded tensor names. This scope is applied to tensor instances loaded into the passed session, but it is *not* written through to the static `MetaGraphDef` protocol buffer that is returned.
IDictionary<string, object> saver_kwargs: Optional keyword arguments passed through to Saver.

Returns

AutoTrackable: The `MetaGraphDef` protocol buffer loaded in the provided session. This can be used to further extract signature-defs, collection-defs, etc.

AutoTrackable load(Session sess, IEnumerable<string> tags, Byte[] export_dir, string import_scope, IDictionary<string, object> saver_kwargs)

Loads the model from a SavedModel as specified by tags. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.loader.load or tf.compat.v1.saved_model.load. There will be a new function for importing SavedModels in Tensorflow 2.0.

Parameters

Session sess: The TensorFlow session to restore the variables.
IEnumerable<string> tags: Set of string tags to identify the required MetaGraphDef. These should correspond to the tags used when saving the variables using the SavedModel `save()` API.
Byte[] export_dir: Directory in which the SavedModel protocol buffer and variables to be loaded are located.
string import_scope: Optional `string` -- if specified, prepend this string followed by '/' to all loaded tensor names. This scope is applied to tensor instances loaded into the passed session, but it is *not* written through to the static `MetaGraphDef` protocol buffer that is returned.
IDictionary<string, object> saver_kwargs: Optional keyword arguments passed through to Saver.

Returns

AutoTrackable: The `MetaGraphDef` protocol buffer loaded in the provided session. This can be used to further extract signature-defs, collection-defs, etc.

AutoTrackable load(_MonitoredSession sess, IEnumerable<string> tags, string export_dir, string import_scope, IDictionary<string, object> saver_kwargs)

Loads the model from a SavedModel as specified by tags. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.loader.load or tf.compat.v1.saved_model.load. There will be a new function for importing SavedModels in Tensorflow 2.0.

Parameters

_MonitoredSession sess: The TensorFlow session to restore the variables.
IEnumerable<string> tags: Set of string tags to identify the required MetaGraphDef. These should correspond to the tags used when saving the variables using the SavedModel `save()` API.
string export_dir: Directory in which the SavedModel protocol buffer and variables to be loaded are located.
string import_scope: Optional `string` -- if specified, prepend this string followed by '/' to all loaded tensor names. This scope is applied to tensor instances loaded into the passed session, but it is *not* written through to the static `MetaGraphDef` protocol buffer that is returned.
IDictionary<string, object> saver_kwargs: Optional keyword arguments passed through to Saver.

Returns

AutoTrackable: The `MetaGraphDef` protocol buffer loaded in the provided session. This can be used to further extract signature-defs, collection-defs, etc.

AutoTrackable load(_MonitoredSession sess, IEnumerable<string> tags, Byte[] export_dir, string import_scope, IDictionary<string, object> saver_kwargs)

Loads the model from a SavedModel as specified by tags. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.loader.load or tf.compat.v1.saved_model.load. There will be a new function for importing SavedModels in Tensorflow 2.0.

Parameters

_MonitoredSession sess: The TensorFlow session to restore the variables.
IEnumerable<string> tags: Set of string tags to identify the required MetaGraphDef. These should correspond to the tags used when saving the variables using the SavedModel `save()` API.
Byte[] export_dir: Directory in which the SavedModel protocol buffer and variables to be loaded are located.
string import_scope: Optional `string` -- if specified, prepend this string followed by '/' to all loaded tensor names. This scope is applied to tensor instances loaded into the passed session, but it is *not* written through to the static `MetaGraphDef` protocol buffer that is returned.
IDictionary<string, object> saver_kwargs: Optional keyword arguments passed through to Saver.

Returns

AutoTrackable: The `MetaGraphDef` protocol buffer loaded in the provided session. This can be used to further extract signature-defs, collection-defs, etc.

AutoTrackable load(LocalCLIDebugWrapperSession sess, IEnumerable<string> tags, string export_dir, string import_scope, IDictionary<string, object> saver_kwargs)

Loads the model from a SavedModel as specified by tags. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.loader.load or tf.compat.v1.saved_model.load. There will be a new function for importing SavedModels in Tensorflow 2.0.

Parameters

LocalCLIDebugWrapperSession sess: The TensorFlow session to restore the variables.
IEnumerable<string> tags: Set of string tags to identify the required MetaGraphDef. These should correspond to the tags used when saving the variables using the SavedModel `save()` API.
string export_dir: Directory in which the SavedModel protocol buffer and variables to be loaded are located.
string import_scope: Optional `string` -- if specified, prepend this string followed by '/' to all loaded tensor names. This scope is applied to tensor instances loaded into the passed session, but it is *not* written through to the static `MetaGraphDef` protocol buffer that is returned.
IDictionary<string, object> saver_kwargs: Optional keyword arguments passed through to Saver.

Returns

AutoTrackable: The `MetaGraphDef` protocol buffer loaded in the provided session. This can be used to further extract signature-defs, collection-defs, etc.

AutoTrackable load(LocalCLIDebugWrapperSession sess, IEnumerable<string> tags, Byte[] export_dir, string import_scope, IDictionary<string, object> saver_kwargs)

Loads the model from a SavedModel as specified by tags. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.loader.load or tf.compat.v1.saved_model.load. There will be a new function for importing SavedModels in Tensorflow 2.0.

Parameters

LocalCLIDebugWrapperSession sess: The TensorFlow session to restore the variables.
IEnumerable<string> tags: Set of string tags to identify the required MetaGraphDef. These should correspond to the tags used when saving the variables using the SavedModel `save()` API.
Byte[] export_dir: Directory in which the SavedModel protocol buffer and variables to be loaded are located.
string import_scope: Optional `string` -- if specified, prepend this string followed by '/' to all loaded tensor names. This scope is applied to tensor instances loaded into the passed session, but it is *not* written through to the static `MetaGraphDef` protocol buffer that is returned.
IDictionary<string, object> saver_kwargs: Optional keyword arguments passed through to Saver.

Returns

AutoTrackable: The `MetaGraphDef` protocol buffer loaded in the provided session. This can be used to further extract signature-defs, collection-defs, etc.

object load_dyn(object sess, object tags, object export_dir, object import_scope, IDictionary<string, object> saver_kwargs)

Loads the model from a SavedModel as specified by tags. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.loader.load or tf.compat.v1.saved_model.load. There will be a new function for importing SavedModels in Tensorflow 2.0.

Parameters

object sess: The TensorFlow session to restore the variables.
object tags: Set of string tags to identify the required MetaGraphDef. These should correspond to the tags used when saving the variables using the SavedModel `save()` API.
object export_dir: Directory in which the SavedModel protocol buffer and variables to be loaded are located.
object import_scope: Optional `string` -- if specified, prepend this string followed by '/' to all loaded tensor names. This scope is applied to tensor instances loaded into the passed session, but it is *not* written through to the static `MetaGraphDef` protocol buffer that is returned.
IDictionary<string, object> saver_kwargs: Optional keyword arguments passed through to Saver.

Returns

object: The `MetaGraphDef` protocol buffer loaded in the provided session. This can be used to further extract signature-defs, collection-defs, etc.

object load_v2(string export_dir, string tags)

Load a SavedModel from `export_dir`.

Signatures associated with the SavedModel are available as functions: Objects exported with tf.saved_model.save additionally have trackable objects and functions assigned to attributes: _Loading Keras models_

Keras models are trackable, so they can be saved to SavedModel. The object returned by tf.saved_model.load is not a Keras object (i.e. doesn't have `.fit`, `.predict`, etc. methods). A few attributes and functions are still available: `.variables`, `.trainable_variables` and `.__call__`. Use tf.keras.models.load_model to restore the Keras model.

_Importing SavedModels from TensorFlow 1.x_

SavedModels from tf.estimator.Estimator or 1.x SavedModel APIs have a flat graph instead of tf.function objects. These SavedModels will have functions corresponding to their signatures in the `.signatures` attribute, but also have a `.prune` method which allows you to extract functions for new subgraphs. This is equivalent to importing the SavedModel and naming feeds and fetches in a Session from TensorFlow 1.x. See `tf.compat.v1.wrap_function` for details. These SavedModels also have a `.variables` attribute containing imported variables, and a `.graph` attribute representing the whole imported graph. For SavedModels exported from tf.saved_model.save, variables are instead assigned to whichever attributes they were assigned before export.

Parameters

string export_dir: The SavedModel directory to load from.
string tags: A tag or sequence of tags identifying the MetaGraph to load. Optional if the SavedModel contains a single MetaGraph, as for those exported from tf.saved_model.load.

Returns

object: A trackable object with a `signatures` attribute mapping from signature keys to functions. If the SavedModel was exported by tf.saved_model.load, it also points to trackable objects and functions which were attached to the exported object.

Show Example

imported = tf.saved_model.load(path)
            f = imported.signatures["serving_default"]
            print(f(x=tf.constant([[1.]])))

object load_v2(string export_dir, IDictionary<string, IEnumerable<int>> tags)

Load a SavedModel from `export_dir`.

Signatures associated with the SavedModel are available as functions: Objects exported with tf.saved_model.save additionally have trackable objects and functions assigned to attributes: _Loading Keras models_

Keras models are trackable, so they can be saved to SavedModel. The object returned by tf.saved_model.load is not a Keras object (i.e. doesn't have `.fit`, `.predict`, etc. methods). A few attributes and functions are still available: `.variables`, `.trainable_variables` and `.__call__`. Use tf.keras.models.load_model to restore the Keras model.

_Importing SavedModels from TensorFlow 1.x_

SavedModels from tf.estimator.Estimator or 1.x SavedModel APIs have a flat graph instead of tf.function objects. These SavedModels will have functions corresponding to their signatures in the `.signatures` attribute, but also have a `.prune` method which allows you to extract functions for new subgraphs. This is equivalent to importing the SavedModel and naming feeds and fetches in a Session from TensorFlow 1.x. See `tf.compat.v1.wrap_function` for details. These SavedModels also have a `.variables` attribute containing imported variables, and a `.graph` attribute representing the whole imported graph. For SavedModels exported from tf.saved_model.save, variables are instead assigned to whichever attributes they were assigned before export.

Parameters

string export_dir: The SavedModel directory to load from.
IDictionary<string, IEnumerable<int>> tags: A tag or sequence of tags identifying the MetaGraph to load. Optional if the SavedModel contains a single MetaGraph, as for those exported from tf.saved_model.load.

Returns

object: A trackable object with a `signatures` attribute mapping from signature keys to functions. If the SavedModel was exported by tf.saved_model.load, it also points to trackable objects and functions which were attached to the exported object.

Show Example

imported = tf.saved_model.load(path)
            f = imported.signatures["serving_default"]
            print(f(x=tf.constant([[1.]])))

object load_v2(Byte[] export_dir, IDictionary<string, IEnumerable<int>> tags)

Load a SavedModel from `export_dir`.

Signatures associated with the SavedModel are available as functions: Objects exported with tf.saved_model.save additionally have trackable objects and functions assigned to attributes: _Loading Keras models_

Keras models are trackable, so they can be saved to SavedModel. The object returned by tf.saved_model.load is not a Keras object (i.e. doesn't have `.fit`, `.predict`, etc. methods). A few attributes and functions are still available: `.variables`, `.trainable_variables` and `.__call__`. Use tf.keras.models.load_model to restore the Keras model.

_Importing SavedModels from TensorFlow 1.x_

SavedModels from tf.estimator.Estimator or 1.x SavedModel APIs have a flat graph instead of tf.function objects. These SavedModels will have functions corresponding to their signatures in the `.signatures` attribute, but also have a `.prune` method which allows you to extract functions for new subgraphs. This is equivalent to importing the SavedModel and naming feeds and fetches in a Session from TensorFlow 1.x. See `tf.compat.v1.wrap_function` for details. These SavedModels also have a `.variables` attribute containing imported variables, and a `.graph` attribute representing the whole imported graph. For SavedModels exported from tf.saved_model.save, variables are instead assigned to whichever attributes they were assigned before export.

Parameters

Byte[] export_dir: The SavedModel directory to load from.
IDictionary<string, IEnumerable<int>> tags: A tag or sequence of tags identifying the MetaGraph to load. Optional if the SavedModel contains a single MetaGraph, as for those exported from tf.saved_model.load.

Returns

object: A trackable object with a `signatures` attribute mapping from signature keys to functions. If the SavedModel was exported by tf.saved_model.load, it also points to trackable objects and functions which were attached to the exported object.

Show Example

imported = tf.saved_model.load(path)
            f = imported.signatures["serving_default"]
            print(f(x=tf.constant([[1.]])))

object load_v2(Byte[] export_dir, IEnumerable<string> tags)

Load a SavedModel from `export_dir`.

Signatures associated with the SavedModel are available as functions: Objects exported with tf.saved_model.save additionally have trackable objects and functions assigned to attributes: _Loading Keras models_

Keras models are trackable, so they can be saved to SavedModel. The object returned by tf.saved_model.load is not a Keras object (i.e. doesn't have `.fit`, `.predict`, etc. methods). A few attributes and functions are still available: `.variables`, `.trainable_variables` and `.__call__`. Use tf.keras.models.load_model to restore the Keras model.

_Importing SavedModels from TensorFlow 1.x_

SavedModels from tf.estimator.Estimator or 1.x SavedModel APIs have a flat graph instead of tf.function objects. These SavedModels will have functions corresponding to their signatures in the `.signatures` attribute, but also have a `.prune` method which allows you to extract functions for new subgraphs. This is equivalent to importing the SavedModel and naming feeds and fetches in a Session from TensorFlow 1.x. See `tf.compat.v1.wrap_function` for details. These SavedModels also have a `.variables` attribute containing imported variables, and a `.graph` attribute representing the whole imported graph. For SavedModels exported from tf.saved_model.save, variables are instead assigned to whichever attributes they were assigned before export.

Parameters

Byte[] export_dir: The SavedModel directory to load from.
IEnumerable<string> tags: A tag or sequence of tags identifying the MetaGraph to load. Optional if the SavedModel contains a single MetaGraph, as for those exported from tf.saved_model.load.

Returns

object: A trackable object with a `signatures` attribute mapping from signature keys to functions. If the SavedModel was exported by tf.saved_model.load, it also points to trackable objects and functions which were attached to the exported object.

Show Example

imported = tf.saved_model.load(path)
            f = imported.signatures["serving_default"]
            print(f(x=tf.constant([[1.]])))

object load_v2(Byte[] export_dir, string tags)

Load a SavedModel from `export_dir`.

Signatures associated with the SavedModel are available as functions: Objects exported with tf.saved_model.save additionally have trackable objects and functions assigned to attributes: _Loading Keras models_

Keras models are trackable, so they can be saved to SavedModel. The object returned by tf.saved_model.load is not a Keras object (i.e. doesn't have `.fit`, `.predict`, etc. methods). A few attributes and functions are still available: `.variables`, `.trainable_variables` and `.__call__`. Use tf.keras.models.load_model to restore the Keras model.

_Importing SavedModels from TensorFlow 1.x_

SavedModels from tf.estimator.Estimator or 1.x SavedModel APIs have a flat graph instead of tf.function objects. These SavedModels will have functions corresponding to their signatures in the `.signatures` attribute, but also have a `.prune` method which allows you to extract functions for new subgraphs. This is equivalent to importing the SavedModel and naming feeds and fetches in a Session from TensorFlow 1.x. See `tf.compat.v1.wrap_function` for details. These SavedModels also have a `.variables` attribute containing imported variables, and a `.graph` attribute representing the whole imported graph. For SavedModels exported from tf.saved_model.save, variables are instead assigned to whichever attributes they were assigned before export.

Parameters

Byte[] export_dir: The SavedModel directory to load from.
string tags: A tag or sequence of tags identifying the MetaGraph to load. Optional if the SavedModel contains a single MetaGraph, as for those exported from tf.saved_model.load.

Returns

object: A trackable object with a `signatures` attribute mapping from signature keys to functions. If the SavedModel was exported by tf.saved_model.load, it also points to trackable objects and functions which were attached to the exported object.

Show Example

imported = tf.saved_model.load(path)
            f = imported.signatures["serving_default"]
            print(f(x=tf.constant([[1.]])))

object load_v2(string export_dir, IEnumerable<string> tags)

Load a SavedModel from `export_dir`.

Signatures associated with the SavedModel are available as functions: Objects exported with tf.saved_model.save additionally have trackable objects and functions assigned to attributes: _Loading Keras models_

Keras models are trackable, so they can be saved to SavedModel. The object returned by tf.saved_model.load is not a Keras object (i.e. doesn't have `.fit`, `.predict`, etc. methods). A few attributes and functions are still available: `.variables`, `.trainable_variables` and `.__call__`. Use tf.keras.models.load_model to restore the Keras model.

_Importing SavedModels from TensorFlow 1.x_

SavedModels from tf.estimator.Estimator or 1.x SavedModel APIs have a flat graph instead of tf.function objects. These SavedModels will have functions corresponding to their signatures in the `.signatures` attribute, but also have a `.prune` method which allows you to extract functions for new subgraphs. This is equivalent to importing the SavedModel and naming feeds and fetches in a Session from TensorFlow 1.x. See `tf.compat.v1.wrap_function` for details. These SavedModels also have a `.variables` attribute containing imported variables, and a `.graph` attribute representing the whole imported graph. For SavedModels exported from tf.saved_model.save, variables are instead assigned to whichever attributes they were assigned before export.

Parameters

string export_dir: The SavedModel directory to load from.
IEnumerable<string> tags: A tag or sequence of tags identifying the MetaGraph to load. Optional if the SavedModel contains a single MetaGraph, as for those exported from tf.saved_model.load.

Returns

object: A trackable object with a `signatures` attribute mapping from signature keys to functions. If the SavedModel was exported by tf.saved_model.load, it also points to trackable objects and functions which were attached to the exported object.

Show Example

imported = tf.saved_model.load(path)
            f = imported.signatures["serving_default"]
            print(f(x=tf.constant([[1.]])))

object load_v2_dyn(object export_dir, object tags)

Load a SavedModel from `export_dir`.

Signatures associated with the SavedModel are available as functions: Objects exported with tf.saved_model.save additionally have trackable objects and functions assigned to attributes: _Loading Keras models_

Keras models are trackable, so they can be saved to SavedModel. The object returned by tf.saved_model.load is not a Keras object (i.e. doesn't have `.fit`, `.predict`, etc. methods). A few attributes and functions are still available: `.variables`, `.trainable_variables` and `.__call__`. Use tf.keras.models.load_model to restore the Keras model.

_Importing SavedModels from TensorFlow 1.x_

SavedModels from tf.estimator.Estimator or 1.x SavedModel APIs have a flat graph instead of tf.function objects. These SavedModels will have functions corresponding to their signatures in the `.signatures` attribute, but also have a `.prune` method which allows you to extract functions for new subgraphs. This is equivalent to importing the SavedModel and naming feeds and fetches in a Session from TensorFlow 1.x. See `tf.compat.v1.wrap_function` for details. These SavedModels also have a `.variables` attribute containing imported variables, and a `.graph` attribute representing the whole imported graph. For SavedModels exported from tf.saved_model.save, variables are instead assigned to whichever attributes they were assigned before export.

Parameters

object export_dir: The SavedModel directory to load from.
object tags: A tag or sequence of tags identifying the MetaGraph to load. Optional if the SavedModel contains a single MetaGraph, as for those exported from tf.saved_model.load.

Returns

object: A trackable object with a `signatures` attribute mapping from signature keys to functions. If the SavedModel was exported by tf.saved_model.load, it also points to trackable objects and functions which were attached to the exported object.

Show Example

imported = tf.saved_model.load(path)
            f = imported.signatures["serving_default"]
            print(f(x=tf.constant([[1.]])))

object main_op_with_restore(object restore_op_name)

Returns a main op to init variables, tables and restore the graph. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.main_op_with_restore or tf.compat.v1.saved_model.main_op.main_op_with_restore.

Returns the main op including the group of ops that initializes all variables, initialize local variables, initialize all tables and the restore op name.

Parameters

object restore_op_name: Name of the op to use to restore the graph.

Returns

object: The set of ops to be run as part of the main op upon the load operation.

object main_op_with_restore_dyn(object restore_op_name)

Returns a main op to init variables, tables and restore the graph. (deprecated)

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.main_op_with_restore or tf.compat.v1.saved_model.main_op.main_op_with_restore.

Returns the main op including the group of ops that initializes all variables, initialize local variables, initialize all tables and the restore op name.

Parameters

object restore_op_name: Name of the op to use to restore the graph.

Returns

object: The set of ops to be run as part of the main op upon the load operation.

object predict_signature_def(IDictionary<string, object> inputs, IDictionary<string, object> outputs)

Creates prediction signature from given inputs and outputs.

This function produces signatures intended for use with the TensorFlow Serving Predict API (tensorflow_serving/apis/prediction_service.proto). This API imposes no constraints on the input and output types.

Parameters

IDictionary<string, object> inputs: dict of string to `Tensor`.
IDictionary<string, object> outputs: dict of string to `Tensor`.

Returns

object: A prediction-flavored signature_def.

object predict_signature_def_dyn(object inputs, object outputs)

Creates prediction signature from given inputs and outputs.

This function produces signatures intended for use with the TensorFlow Serving Predict API (tensorflow_serving/apis/prediction_service.proto). This API imposes no constraints on the input and output types.

Parameters

object inputs: dict of string to `Tensor`.
object outputs: dict of string to `Tensor`.

Returns

object: A prediction-flavored signature_def.

object regression_signature_def(IGraphNodeBase examples, IGraphNodeBase predictions)

Creates regression signature from given examples and predictions.

This function produces signatures intended for use with the TensorFlow Serving Regress API (tensorflow_serving/apis/prediction_service.proto), and so constrains the input and output types to those allowed by TensorFlow Serving.

Parameters

IGraphNodeBase examples: A string `Tensor`, expected to accept serialized tf.Examples.
IGraphNodeBase predictions: A float `Tensor`.

Returns

object: A regression-flavored signature_def.

object regression_signature_def_dyn(object examples, object predictions)

Creates regression signature from given examples and predictions.

This function produces signatures intended for use with the TensorFlow Serving Regress API (tensorflow_serving/apis/prediction_service.proto), and so constrains the input and output types to those allowed by TensorFlow Serving.

Parameters

object examples: A string `Tensor`, expected to accept serialized tf.Examples.
object predictions: A float `Tensor`.

Returns

object: A regression-flavored signature_def.

void save(object obj, IGraphNodeBase export_dir, object signatures)

Exports the Trackable object `obj` to [SavedModel format](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/saved_model/README.md).

Example usage: The resulting SavedModel is then servable with an input named "x", its value having any shape and dtype float32.

The optional `signatures` argument controls which methods in `obj` will be available to programs which consume `SavedModel`s, for example serving APIs. Python functions may be decorated with `@tf.function(input_signature=...)` and passed as signatures directly, or lazily with a call to `get_concrete_function` on the method decorated with `@tf.function`.

If the `signatures` argument is omitted, `obj` will be searched for `@tf.function`-decorated methods. If exactly one `@tf.function` is found, that method will be used as the default signature for the SavedModel. This behavior is expected to change in the future, when a corresponding tf.saved_model.load symbol is added. At that point signatures will be completely optional, and any `@tf.function` attached to `obj` or its dependencies will be exported for use with `load`.

When invoking a signature in an exported SavedModel, `Tensor` arguments are identified by name. These names will come from the Python function's argument names by default. They may be overridden by specifying a `name=...` argument in the corresponding tf.TensorSpec object. Explicit naming is required if multiple `Tensor`s are passed through a single argument to the Python function.

The outputs of functions used as `signatures` must either be flat lists, in which case outputs will be numbered, or a dictionary mapping string keys to `Tensor`, in which case the keys will be used to name outputs.

Signatures are available in objects returned by tf.saved_model.load as a `.signatures` attribute. This is a reserved attribute: tf.saved_model.save on an object with a custom `.signatures` attribute will raise an exception.

Since tf.keras.Model objects are also Trackable, this function can be used to export Keras models. For example, exporting with a signature specified: Exporting from a function without a fixed signature: tf.keras.Model instances constructed from inputs and outputs already have a signature and so do not require a `@tf.function` decorator or a `signatures` argument. If neither are specified, the model's forward pass is exported. Variables must be tracked by assigning them to an attribute of a tracked object or to an attribute of `obj` directly. TensorFlow objects (e.g. layers from tf.keras.layers, optimizers from tf.train) track their variables automatically. This is the same tracking scheme that tf.train.Checkpoint uses, and an exported `Checkpoint` object may be restored as a training checkpoint by pointing tf.train.Checkpoint.restore to the SavedModel's "variables/" subdirectory. Currently variables are the only stateful objects supported by tf.saved_model.save, but others (e.g. tables) will be supported in the future.

tf.function does not hard-code device annotations from outside the function body, instead using the calling context's device. This means for example that exporting a model which runs on a GPU and serving it on a CPU will generally work, with some exceptions. tf.device annotations inside the body of the function will be hard-coded in the exported model; this type of annotation is discouraged. Device-specific operations, e.g. with "cuDNN" in the name or with device-specific layouts, may cause issues. Currently a `DistributionStrategy` is another exception: active distribution strategies will cause device placements to be hard-coded in a function. Exporting a single-device computation and importing under a `DistributionStrategy` is not currently supported, but may be in the future.

SavedModels exported with tf.saved_model.save [strip default-valued attributes](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/python/saved_model/README.md#stripping-default-valued-attributes) automatically, which removes one source of incompatibilities when the consumer of a SavedModel is running an older TensorFlow version than the producer. There are however other sources of incompatibilities which are not handled automatically, such as when the exported model contains operations which the consumer does not have definitions for.

Parameters

object obj: A trackable object to export.
IGraphNodeBase export_dir: A directory in which to write the SavedModel.
object signatures: Optional, either a tf.function with an input signature specified or the result of `f.get_concrete_function` on a `@tf.function`-decorated function `f`, in which case `f` will be used to generate a signature for the SavedModel under the default serving signature key. `signatures` may also be a dictionary, in which case it maps from signature keys to either tf.function instances with input signatures or concrete functions. The keys of such a dictionary may be arbitrary strings, but will typically be from the tf.saved_model.signature_constants module.

Show Example

class Adder(tf.Module):  @tf.function(input_signature=[tf.TensorSpec(shape=None, dtype=tf.float32)])
def add(self, x):
  return x + x + 1.  to_export = Adder()
tf.saved_model.save(to_export, '/tmp/adder')

void save(object obj, Byte[] export_dir, IDictionary<string, object> signatures)