tensorflow::serving::TfrtSavedModelFactory Class Reference

#include <tfrt_saved_model_factory.h>

Public Types
using	Batcher = SharedBatchScheduler< SavedModelBatchingTask >

Public Member Functions
	TfrtSavedModelFactory (const TfrtSavedModelConfig &config, std::shared_ptr< Batcher > batch_scheduler, std::unique_ptr< ThreadPoolFactory > thread_pool_factory)
	TfrtSavedModelFactory (const TfrtSavedModelConfig &config, std::shared_ptr< Batcher > batch_scheduler, std::unique_ptr< ThreadPoolFactory > thread_pool_factory, std::function< std::unique_ptr< RequestRecorder >(TfrtSavedModelServable &)> recorder_creator)
virtual absl::Status	CreateTfrtSavedModelWithMetadata (const Loader::Metadata &metadata, const string &path, std::unique_ptr< Servable > *servable)
absl::Status	CreateTfrtSavedModelWithMetadata (const Loader::Metadata &metadata, const string &path, std::unique_ptr< tfrt_stub::SavedModel > *saved_model)
absl::Status	EstimateResourceRequirement (const string &path, ResourceAllocation *estimate) const
const TfrtSavedModelConfig &	config () const
TfrtSavedModelConfig &	mutable_config ()
absl::string_view	GetServingResourceType () const

Static Public Member Functions
static absl::Status	Create (const TfrtSavedModelConfig &config, std::unique_ptr< TfrtSavedModelFactory > *factory)

Detailed Description

A factory that creates tfrt_stub::SavedModel from SavedModel export paths.

The factory can also estimate the resource (e.g. RAM) requirements of a tfrt_stub::SavedModel based on the SavedModel (i.e. prior to loading the session).

This class is thread-safe.

Definition at line 49 of file tfrt_saved_model_factory.h.

Member Function Documentation

Create()

absl::Status tensorflow::serving::TfrtSavedModelFactory::Create ( const TfrtSavedModelConfig &  config, std::unique_ptr< TfrtSavedModelFactory > *  factory  )

static

Instantiates a TfrtSavedModelFactory using a config.

Parameters

config	Config with initialization options.
factory	Newly created factory if the returned Status is OK.

Definition at line 152 of file tfrt_saved_model_factory.cc.

CreateTfrtSavedModelWithMetadata()

virtual absl::Status tensorflow::serving::TfrtSavedModelFactory::CreateTfrtSavedModelWithMetadata ( const Loader::Metadata &  metadata, const string &  path, std::unique_ptr< Servable > *  servable  )

virtual

Instantiates a tfrt_stub::SavedModel from a given export or SavedModel path and the given metadata.

Parameters

metadata	Metadata to be associated with the saved_model.
path	Path to the model.
servable	Newly created Servable if the returned Status is OK.

EstimateResourceRequirement()

absl::Status tensorflow::serving::TfrtSavedModelFactory::EstimateResourceRequirement	(	const string &	path,
		ResourceAllocation *	estimate
	)		const

Estimates the resources a SavedModel will use once loaded, from its export path.

Parameters

path	Path to the model.
estimate	Output resource usage estimates. Different kinds of resources (e.g. CPU, RAM, etc.) may get populated.

Definition at line 164 of file tfrt_saved_model_factory.cc.

---

The documentation for this class was generated from the following files:

• tensorflow_serving/servables/tensorflow/tfrt_saved_model_factory.h
• tensorflow_serving/servables/tensorflow/tfrt_saved_model_factory.cc