tensorflow::serving

Summary

Typedefs

LoaderSource using
Source< std::unique_ptr< Loader >>

Functions

operator!=(const Loader::Metadata & a, const Loader::Metadata & b)
bool
operator!=(const ServableHandle< T > & l, const ServableHandle< U > & r)
constexpr bool
operator!=(const ServableStateMonitor::ServableStateAndTime & a, const ServableStateMonitor::ServableStateAndTime & b)
bool
operator<<(std::ostream & os, const ServableStateMonitor::ServableStateAndTime & state_and_time)
std::ostream &
operator==(const AspiredVersionPolicy::ServableAction & lhs, const AspiredVersionPolicy::ServableAction & rhs)
bool
operator==(const Loader::Metadata & a, const Loader::Metadata & b)
bool
operator==(const ServableHandle< T > & l, const ServableHandle< U > & r)
constexpr bool
operator==(const ServableStateMonitor::ServableStateAndTime & a, const ServableStateMonitor::ServableStateAndTime & b)
bool

Classes

tensorflow::serving::AspiredVersionPolicy

An interface for the policy to be applied for transitioning servable versions in a servable stream.

tensorflow::serving::AspiredVersionsManager

A manager that implements the Target<Loader> API which uses aspired-versions callbacks to dictate which servable versions to load.

tensorflow::serving::AspiredVersionsManagerBuilder

Builds an AspiredVersionsManager with options and sources connected to it.

tensorflow::serving::BasicManager

Helps manage the lifecycle of servables including loading, serving and unloading them.

tensorflow::serving::CachingManager

A manager that manages and loads servables on-demand.

tensorflow::serving::ClassifierInterface

Model-type agnostic interface for performing classification.

tensorflow::serving::FileSystemStoragePathSource

A storage path source that aspires versions for a given set of servables.

tensorflow::serving::Loader

A standardized abstraction for an object that manages the lifecycle of a servable, including loading and unloading it.

tensorflow::serving::LoaderHarness

LoaderHarness is a widget the Manager uses to hold on to and talk to a Loader while it owns it.

tensorflow::serving::Manager

Manager is responsible for loading, unloading, lookup and lifetime management of all Servable objects via their Loaders.

tensorflow::serving::PathPrefixLoaderFactory

A simple LoaderFactory that looks for a servable at a path formed by concatenating a fixed path prefix with the servable's name.

tensorflow::serving::RegressorInterface

Model agnostic interface for performing regression.

tensorflow::serving::ResourceUnsafeLoader

A Loader that is oblivious to resources.

tensorflow::serving::SavedModelBundleFactory

A factory that creates SavedModelBundles from SavedModel or SessionBundle export paths.

tensorflow::serving::ServableHandle

A smart pointer to the underlying servable object T retrieved from the Loader.

tensorflow::serving::ServableStateMonitor

A utility that listens to an EventBus<ServableState>, and keeps track of the state of each servable mentioned on the bus.

tensorflow::serving::ServerCore

ServerCore contains state and helper methods enabling the building of ModelServers that support multiple interfaces.

tensorflow::serving::ServingSession

A Session that blocks state-changing methods such as Close(), while allowing Run() for read-only access (not enforced).

tensorflow::serving::ServingSessionWrapper

A ServingSession that wraps a given Session, and blocks all calls other than Run().

tensorflow::serving::SessionWrapperIgnoreThreadPoolOptions
tensorflow::serving::SharedPtrHandle

An implementation of UntypedServableHandle using shared_ptr to do ref-counting on the Loader that owns the Servable.

tensorflow::serving::Source

An abstraction for a module that sources servables to load, or, more precisely, handles to data that can be used to load those servables.

tensorflow::serving::SourceAdapter

An abstraction for a module that receives aspired-version callbacks with data of type InputType and converts them into calls with data of type OutputType.

tensorflow::serving::UntypedServableHandle

A non-templatized handle to a servable, used internally in the Manager to retrieve a type-erased servable object from the Loader.

Structs

tensorflow::serving::AspiredServableStateSnapshot

A snapshot of a servable's state and aspiredness.

tensorflow::serving::ServableRequest

A query for a specific loaded servable object.

Namespaces

tensorflow::serving::internal
tensorflow::serving::test_util

Typedefs

LoaderSource

Source< std::unique_ptr< Loader >> LoaderSource

Functions

operator!=

bool operator!=(
  const Loader::Metadata & a,
  const Loader::Metadata & b
)

operator!=

constexpr bool operator!=(
  const ServableHandle< T > & l,
  const ServableHandle< U > & r
)

operator!=

bool operator!=(
  const ServableStateMonitor::ServableStateAndTime & a,
  const ServableStateMonitor::ServableStateAndTime & b
)

operator<<

std::ostream & operator<<(
  std::ostream & os,
  const ServableStateMonitor::ServableStateAndTime & state_and_time
)

operator==

bool operator==(
  const AspiredVersionPolicy::ServableAction & lhs,
  const AspiredVersionPolicy::ServableAction & rhs
)

operator==

bool operator==(
  const Loader::Metadata & a,
  const Loader::Metadata & b
)

operator==

constexpr bool operator==(
  const ServableHandle< T > & l,
  const ServableHandle< U > & r
)

operator==

bool operator==(
  const ServableStateMonitor::ServableStateAndTime & a,
  const ServableStateMonitor::ServableStateAndTime & b
)