FAdamOptimizer | Unreal Engine 5.7 Documentation

Adam Optimizer

Estimates gradient via finite differencing and does gradient descent. Roughly based on https://arxiv.org/abs/1910.06513


Name	FAdamOptimizer
Type	struct
Header File	/Engine/Plugins/Experimental/LearningAgents/Source/Learning/Public/LearningAdamOptimizer.h
Include Path	#include "LearningAdamOptimizer.h"

Syntax

struct FAdamOptimizer : public UE::Learning::IOptimizer

Name	Remarks	Include Path	Unreal Specifiers
FAdamOptimizer ( const uint32 Seed, const FAdamOptimizerSettings& InSettings )		LearningAdamOptimizer.h

Name	Type	Include Path
Estimate	TLearningArray< 1, float >	LearningAdamOptimizer.h
GaussianSamples	TLearningArray< 2, float >	LearningAdamOptimizer.h
Gradient	TLearningArray< 1, float >	LearningAdamOptimizer.h
Iterations	TLearningArray< 1, int32 >	LearningAdamOptimizer.h
M0	TLearningArray< 1, float >	LearningAdamOptimizer.h
M1	TLearningArray< 1, float >	LearningAdamOptimizer.h
M1HatMax	TLearningArray< 1, float >	LearningAdamOptimizer.h
Seed	uint32	LearningAdamOptimizer.h
Settings	FAdamOptimizerSettings	LearningAdamOptimizer.h

Name	Remarks	Include Path
virtual void Reset ( TLearningArrayView< 2, float > OutSamples, const TLearningArrayView< 1, const float > InitialGuess )	Reset the Optimizer	LearningAdamOptimizer.h
virtual void Resize ( const int32 SampleNum, const int32 DimensionsNum )	Resize the Optimizer	LearningAdamOptimizer.h
virtual void Update ( TLearningArrayView< 2, float > InOutSamples, const TLearningArrayView< 1, const float > Losses, const ELogSetting LogSettings )	Update the Optimizer	LearningAdamOptimizer.h