oa::SacTrainer

Minimal fixed-alpha SAC trainer. Actor forward returns [B, 2*A] containing mean then log-standard-deviation; each critic consumes [observation, action] concatenated on the last axis and returns [B] or [B,1]. Its independent actor and critic update loops are composed rather than hidden by inheritance.

ClC++ public headers

Public Methods

const oa::ItTraining & oa::SacTrainer::actorTrainingLoop()

oa::ItTraining & oa::SacTrainer::actorTrainingLoop()

bool oa::SacTrainer::isDone()

const SacTrainerMetrics & oa::SacTrainer::metrics()

oa::Status oa::SacTrainer::syncTargets()

const oa::ItTraining & oa::SacTrainer::trainingLoop()

oa::ItTraining & oa::SacTrainer::trainingLoop()

oa::Status oa::SacTrainer::update()

Static Public Methods

static oa::Result<oa::UniquePtr<SacTrainer>> oa::SacTrainer::create(oa::Engine & inEngine, oa::Module & inActor, oa::Module & inCritic1, oa::Module & inCritic2, oa::Module & inTargetCritic1, oa::Module & inTargetCritic2, oa::Optimizer & inActorOptimizer, oa::Optimizer & inCriticOptimizer, ReplayBuffer & inReplay, const SacTrainerConfig & inConfig)

Constructor & Destructor Documentation

void oa::SacTrainer::~SacTrainer()

No public source comment is attached to this declaration.

Public Method Documentation

const oa::ItTraining & oa::SacTrainer::actorTrainingLoop()

No public source comment is attached to this declaration.

Returns

const oa::ItTraining &

The declared return value.

oa::ItTraining & oa::SacTrainer::actorTrainingLoop()

No public source comment is attached to this declaration.

Returns

oa::ItTraining &

The declared return value.

bool oa::SacTrainer::isDone()

No public source comment is attached to this declaration.

Returns

bool

The declared return value.

const SacTrainerMetrics & oa::SacTrainer::metrics()

No public source comment is attached to this declaration.

Returns

const SacTrainerMetrics &

The declared return value.

oa::Status oa::SacTrainer::syncTargets()

No public source comment is attached to this declaration.

Returns

oa::Status

The declared return value.

const oa::ItTraining & oa::SacTrainer::trainingLoop()

No public source comment is attached to this declaration.

Returns

const oa::ItTraining &

The declared return value.

oa::ItTraining & oa::SacTrainer::trainingLoop()

SAC has two exact optimizer units. The critic loop is the primary update controller; the actor loop remains separately observable.

Returns

oa::ItTraining &

The declared return value.

oa::Status oa::SacTrainer::update()

No public source comment is attached to this declaration.

Returns

oa::Status

The declared return value.

static oa::Result<oa::UniquePtr<SacTrainer>> oa::SacTrainer::create(
	oa::Engine & inEngine,
	oa::Module & inActor,
	oa::Module & inCritic1,
	oa::Module & inCritic2,
	oa::Module & inTargetCritic1,
	oa::Module & inTargetCritic2,
	oa::Optimizer & inActorOptimizer,
	oa::Optimizer & inCriticOptimizer,
	ReplayBuffer & inReplay,
	const SacTrainerConfig & inConfig
)

No public source comment is attached to this declaration.

Parameters

inEngine

oa::Engine &

—

inActor

oa::Module &

—

inCritic1

oa::Module &

—

inCritic2

oa::Module &

—

inTargetCritic1

oa::Module &

—

inTargetCritic2

oa::Module &

—

inActorOptimizer

oa::Optimizer &

—

inCriticOptimizer

oa::Optimizer &

—

inReplay

ReplayBuffer &

—

inConfig

const SacTrainerConfig &

—

Returns

oa::Result<oa::UniquePtr<SacTrainer>>

The declared return value.