Package apple.metalperformanceshaders

Class MPSNNOptimizerStochasticGradientDescent

java.lang.Object

org.moe.natj.general.NativeObject

org.moe.natj.objc.ObjCObject

apple.NSObject

apple.metalperformanceshaders.MPSKernel

apple.metalperformanceshaders.MPSNNOptimizer

apple.metalperformanceshaders.MPSNNOptimizerStochasticGradientDescent

All Implemented Interfaces:

NSCoding, NSCopying, NSSecureCoding, NSObject

public class MPSNNOptimizerStochasticGradientDescent
extends MPSNNOptimizer

MPSNNOptimizerStochasticGradientDescent The MPSNNOptimizerStochasticGradientDescent performs a gradient descent with an optional momentum Update RMSProp is also known as root mean square propagation. useNesterov == NO: m[t] = momentumScale * m[t-1] + learningRate * g variable = variable - m[t] useNesterov == YES: m[t] = momentumScale * m[t-1] + g variable = variable - (learningRate * (g + m[t] * momentumScale)) where, g is gradient of error wrt variable m[t] is momentum of gradients it is a state we keep updating every update iteration API-Since: 12.0

Nested Class Summary

Nested classes/interfaces inherited from class apple.NSObject

NSObject.Block_accessibilityActivateBlock_ret, NSObject.Block_accessibilityActivationPointBlock_ret, NSObject.Block_accessibilityAttributedHintBlock_ret, NSObject.Block_accessibilityAttributedLabelBlock_ret, NSObject.Block_accessibilityAttributedUserInputLabelsBlock_ret, NSObject.Block_accessibilityAttributedValueBlock_ret, NSObject.Block_accessibilityContainerTypeBlock_ret, NSObject.Block_accessibilityCustomActionsBlock_ret, NSObject.Block_accessibilityCustomRotorsBlock_ret, NSObject.Block_accessibilityDecrementBlock_ret, NSObject.Block_accessibilityElementsBlock_ret, NSObject.Block_accessibilityElementsHiddenBlock_ret, NSObject.Block_accessibilityFrameBlock_ret, NSObject.Block_accessibilityHeaderElementsBlock_ret, NSObject.Block_accessibilityHintBlock_ret, NSObject.Block_accessibilityIdentifierBlock_ret, NSObject.Block_accessibilityIncrementBlock_ret, NSObject.Block_accessibilityLabelBlock_ret, NSObject.Block_accessibilityLanguageBlock_ret, NSObject.Block_accessibilityMagicTapBlock_ret, NSObject.Block_accessibilityNavigationStyleBlock_ret, NSObject.Block_accessibilityPathBlock_ret, NSObject.Block_accessibilityPerformEscapeBlock_ret, NSObject.Block_accessibilityRespondsToUserInteractionBlock_ret, NSObject.Block_accessibilityShouldGroupAccessibilityChildrenBlock_ret, NSObject.Block_accessibilityTextualContextBlock_ret, NSObject.Block_accessibilityTraitsBlock_ret, NSObject.Block_accessibilityUserInputLabelsBlock_ret, NSObject.Block_accessibilityValueBlock_ret, NSObject.Block_accessibilityViewIsModalBlock_ret, NSObject.Block_automationElementsBlock_ret, NSObject.Block_isAccessibilityElementBlock_ret, NSObject.Block_setAccessibilityActivateBlock, NSObject.Block_setAccessibilityActivationPointBlock, NSObject.Block_setAccessibilityAttributedHintBlock, NSObject.Block_setAccessibilityAttributedLabelBlock, NSObject.Block_setAccessibilityAttributedUserInputLabelsBlock, NSObject.Block_setAccessibilityAttributedValueBlock, NSObject.Block_setAccessibilityContainerTypeBlock, NSObject.Block_setAccessibilityCustomActionsBlock, NSObject.Block_setAccessibilityCustomRotorsBlock, NSObject.Block_setAccessibilityDecrementBlock, NSObject.Block_setAccessibilityElementsBlock, NSObject.Block_setAccessibilityElementsHiddenBlock, NSObject.Block_setAccessibilityFrameBlock, NSObject.Block_setAccessibilityHeaderElementsBlock, NSObject.Block_setAccessibilityHintBlock, NSObject.Block_setAccessibilityIdentifierBlock, NSObject.Block_setAccessibilityIncrementBlock, NSObject.Block_setAccessibilityLabelBlock, NSObject.Block_setAccessibilityLanguageBlock, NSObject.Block_setAccessibilityMagicTapBlock, NSObject.Block_setAccessibilityNavigationStyleBlock, NSObject.Block_setAccessibilityPathBlock, NSObject.Block_setAccessibilityPerformEscapeBlock, NSObject.Block_setAccessibilityRespondsToUserInteractionBlock, NSObject.Block_setAccessibilityShouldGroupAccessibilityChildrenBlock, NSObject.Block_setAccessibilityTextualContextBlock, NSObject.Block_setAccessibilityTraitsBlock, NSObject.Block_setAccessibilityUserInputLabelsBlock, NSObject.Block_setAccessibilityValueBlock, NSObject.Block_setAccessibilityViewIsModalBlock, NSObject.Block_setAutomationElementsBlock, NSObject.Block_setIsAccessibilityElementBlock, NSObject.Function_instanceMethodForSelector_ret, NSObject.Function_methodForSelector_ret

Constructor Summary

Constructors

Modifier	Constructor	Description
protected	MPSNNOptimizerStochasticGradientDescent(org.moe.natj.general.Pointer peer)

Method Summary

Modifier and Type	Method	Description
boolean	_supportsSecureCoding()	This property must return YES on all classes that allow secure coding.
static boolean	accessInstanceVariablesDirectly()
static MPSNNOptimizerStochasticGradientDescent	alloc()
static MPSNNOptimizerStochasticGradientDescent	allocWithZone(org.moe.natj.general.ptr.VoidPtr zone)
static boolean	automaticallyNotifiesObserversForKey(@NotNull java.lang.String key)
static void	cancelPreviousPerformRequestsWithTarget(@NotNull java.lang.Object aTarget)
static void	cancelPreviousPerformRequestsWithTargetSelectorObject(@NotNull java.lang.Object aTarget, @NotNull org.moe.natj.objc.SEL aSelector, @Nullable java.lang.Object anArgument)
static @NotNull NSArray<java.lang.String>	classFallbacksForKeyedArchiver()
static @NotNull org.moe.natj.objc.Class	classForKeyedUnarchiver()
static java.lang.String	debugDescription_static()
static java.lang.String	description_static()
void	encodeToCommandBufferBatchNormalizationGradientStateBatchNormalizationSourceStateInputMomentumVectorsResultState(@NotNull MTLCommandBuffer commandBuffer, @NotNull MPSCNNBatchNormalizationState batchNormalizationGradientState, @NotNull MPSCNNBatchNormalizationState batchNormalizationSourceState, @Nullable NSArray<? extends MPSVector> inputMomentumVectors, @NotNull MPSCNNNormalizationGammaAndBetaState resultState)	Encode an MPSNNOptimizerStochasticGradientDescent object to a command buffer to perform out of place update The following operations would be applied useNesterov == NO: m[t] = momentumScale * m[t-1] + learningRate * g variable = variable - m[t] useNesterov == YES: m[t] = momentumScale * m[t-1] + g variable = variable - (learningRate * (g + m[t] * momentumScale)) inputMomentumVector == nil variable = variable - (learningRate * g) where, g is gradient of error wrt variable m[t] is momentum of gradients it is a state we keep updating every update iteration
void	encodeToCommandBufferBatchNormalizationStateInputMomentumVectorsResultState(@NotNull MTLCommandBuffer commandBuffer, @NotNull MPSCNNBatchNormalizationState batchNormalizationState, @Nullable NSArray<? extends MPSVector> inputMomentumVectors, @NotNull MPSCNNNormalizationGammaAndBetaState resultState)	Encode an MPSNNOptimizerStochasticGradientDescent object to a command buffer to perform out of place update The following operations would be applied useNesterov == NO: m[t] = momentumScale * m[t-1] + learningRate * g variable = variable - m[t] useNesterov == YES: m[t] = momentumScale * m[t-1] + g variable = variable - (learningRate * (g + m[t] * momentumScale)) inputMomentumVector == nil variable = variable - (learningRate * g) where, g is gradient of error wrt variable m[t] is momentum of gradients it is a state we keep updating every update iteration
void	encodeToCommandBufferConvolutionGradientStateConvolutionSourceStateInputMomentumVectorsResultState(@NotNull MTLCommandBuffer commandBuffer, @NotNull MPSCNNConvolutionGradientState convolutionGradientState, @NotNull MPSCNNConvolutionWeightsAndBiasesState convolutionSourceState, @Nullable NSArray<? extends MPSVector> inputMomentumVectors, @NotNull MPSCNNConvolutionWeightsAndBiasesState resultState)	Encode an MPSNNOptimizerStochasticGradientDescent object to a command buffer to perform out of place update The following operations would be applied useNesterov == NO: m[t] = momentumScale * m[t-1] + learningRate * g variable = variable - m[t] useNesterov == YES: m[t] = momentumScale * m[t-1] + g variable = variable - (learningRate * (g + m[t] * momentumScale)) inputMomentumVector == nil variable = variable - (learningRate * g) where, g is gradient of error wrt variable m[t] is momentum of gradients it is a state we keep updating every update iteration
void	encodeToCommandBufferInputGradientMatrixInputValuesMatrixInputMomentumMatrixResultValuesMatrix(@NotNull MTLCommandBuffer commandBuffer, @NotNull MPSMatrix inputGradientMatrix, @NotNull MPSMatrix inputValuesMatrix, @Nullable MPSMatrix inputMomentumMatrix, @NotNull MPSMatrix resultValuesMatrix)	API-Since: 13.0
void	encodeToCommandBufferInputGradientVectorInputValuesVectorInputMomentumVectorResultValuesVector(@NotNull MTLCommandBuffer commandBuffer, @NotNull MPSVector inputGradientVector, @NotNull MPSVector inputValuesVector, @Nullable MPSVector inputMomentumVector, @NotNull MPSVector resultValuesVector)	Encode an MPSNNOptimizerStochasticGradientDescent object to a command buffer to perform out of place update The following operations would be applied useNesterov == NO: m[t] = momentumScale * m[t-1] + learningRate * g variable = variable - m[t] useNesterov == YES: m[t] = momentumScale * m[t-1] + g variable = variable - (learningRate * (g + m[t] * momentumScale)) inputMomentumVector == nil variable = variable - (learningRate * g) where, g is gradient of error wrt variable m[t] is momentum of gradients it is a state we keep updating every update iteration
static long	hash_static()
MPSNNOptimizerStochasticGradientDescent	init()
MPSNNOptimizerStochasticGradientDescent	initWithCoder(@NotNull NSCoder aDecoder)	NS_DESIGNATED_INITIALIZER
MPSNNOptimizerStochasticGradientDescent	initWithCoderDevice(@NotNull NSCoder aDecoder, @NotNull java.lang.Object device)	NSSecureCoding compatability While the standard NSSecureCoding/NSCoding method -initWithCoder: should work, since the file can't know which device your data is allocated on, we have to guess and may guess incorrectly.
MPSNNOptimizerStochasticGradientDescent	initWithDevice(@NotNull java.lang.Object device)	Standard init with default properties per filter type
MPSNNOptimizerStochasticGradientDescent	initWithDeviceLearningRate(@NotNull MTLDevice device, float learningRate)	Convenience initialization for the momentum update
MPSNNOptimizerStochasticGradientDescent	initWithDeviceMomentumScaleUseNesterovMomentumOptimizerDescriptor(@NotNull MTLDevice device, float momentumScale, boolean useNesterovMomentum, @NotNull MPSNNOptimizerDescriptor optimizerDescriptor)	Full initialization for the momentum update
MPSNNOptimizerStochasticGradientDescent	initWithDeviceMomentumScaleUseNestrovMomentumOptimizerDescriptor(@NotNull MTLDevice device, float momentumScale, boolean useNestrovMomentum, @NotNull MPSNNOptimizerDescriptor optimizerDescriptor)
static NSObject.Function_instanceMethodForSelector_ret	instanceMethodForSelector(org.moe.natj.objc.SEL aSelector)
static NSMethodSignature	instanceMethodSignatureForSelector(org.moe.natj.objc.SEL aSelector)
static boolean	instancesRespondToSelector(org.moe.natj.objc.SEL aSelector)
static boolean	isSubclassOfClass(org.moe.natj.objc.Class aClass)
static @NotNull NSSet<java.lang.String>	keyPathsForValuesAffectingValueForKey(@NotNull java.lang.String key)
float	momentumScale()	[@property] momentumScale The momentumScale at which we update momentum for values array Default value is 0.0
static MPSNNOptimizerStochasticGradientDescent	new_objc()
static boolean	resolveClassMethod(org.moe.natj.objc.SEL sel)
static boolean	resolveInstanceMethod(org.moe.natj.objc.SEL sel)
static void	setVersion_static(long aVersion)
static org.moe.natj.objc.Class	superclass_static()
static boolean	supportsSecureCoding()
boolean	useNesterovMomentum()	[@property] useNesterovMomentum Nesterov momentum is considered an improvement on the usual momentum update Default value is NO [@note] Maps to old useNestrovMomentum property API-Since: 14.0
boolean	useNestrovMomentum()
static boolean	useStoredAccessor()	Deprecated.
static long	version_static()

Methods inherited from class apple.metalperformanceshaders.MPSNNOptimizer

applyGradientClipping, gradientClipMax, gradientClipMin, gradientRescale, learningRate, regularizationScale, regularizationType, setApplyGradientClipping, setLearningRate

Methods inherited from class apple.metalperformanceshaders.MPSKernel

copyWithZone, copyWithZoneDevice, device, encodeWithCoder, label, options, setLabel, setOptions

Methods inherited from class apple.NSObject

accessibilityActivate, accessibilityActivateBlock, accessibilityActivationPoint, accessibilityActivationPointBlock, accessibilityAssistiveTechnologyFocusedIdentifiers, accessibilityAttributedHint, accessibilityAttributedHintBlock, accessibilityAttributedLabel, accessibilityAttributedLabelBlock, accessibilityAttributedUserInputLabels, accessibilityAttributedUserInputLabelsBlock, accessibilityAttributedValue, accessibilityAttributedValueBlock, accessibilityContainerType, accessibilityContainerTypeBlock, accessibilityCustomActions, accessibilityCustomActionsBlock, accessibilityCustomRotors, accessibilityCustomRotorsBlock, accessibilityDecrement, accessibilityDecrementBlock, accessibilityDirectTouchOptions, accessibilityDragSourceDescriptors, accessibilityDropPointDescriptors, accessibilityElementAtIndex, accessibilityElementCount, accessibilityElementDidBecomeFocused, accessibilityElementDidLoseFocus, accessibilityElementIsFocused, accessibilityElements, accessibilityElementsBlock, accessibilityElementsHidden, accessibilityElementsHiddenBlock, accessibilityFrame, accessibilityFrameBlock, accessibilityHeaderElementsBlock, accessibilityHint, accessibilityHintBlock, accessibilityIdentifierBlock, accessibilityIncrement, accessibilityIncrementBlock, accessibilityLabel, accessibilityLabelBlock, accessibilityLanguage, accessibilityLanguageBlock, accessibilityMagicTapBlock, accessibilityNavigationStyle, accessibilityNavigationStyleBlock, accessibilityPath, accessibilityPathBlock, accessibilityPerformEscape, accessibilityPerformEscapeBlock, accessibilityPerformMagicTap, accessibilityRespondsToUserInteraction, accessibilityRespondsToUserInteractionBlock, accessibilityScroll, accessibilityShouldGroupAccessibilityChildrenBlock, accessibilityTextualContext, accessibilityTextualContextBlock, accessibilityTraits, accessibilityTraitsBlock, accessibilityUserInputLabels, accessibilityUserInputLabelsBlock, accessibilityValue, accessibilityValueBlock, accessibilityViewIsModal, accessibilityViewIsModalBlock, accessibilityZoomInAtPoint, accessibilityZoomOutAtPoint, addObserverForKeyPathOptionsContext, attemptRecoveryFromErrorOptionIndex, attemptRecoveryFromErrorOptionIndexDelegateDidRecoverSelectorContextInfo, autoContentAccessingProxy, automationElements, automationElementsBlock, awakeAfterUsingCoder, awakeFromNib, class_objc, classForCoder, classForKeyedArchiver, copy, dealloc, debugDescription, description, dictionaryWithValuesForKeys, didChangeValueForKey, didChangeValueForKeyWithSetMutationUsingObjects, didChangeValuesAtIndexesForKey, doesNotRecognizeSelector, fileManagerShouldProceedAfterError, fileManagerWillProcessPath, finalize_objc, forwardingTargetForSelector, forwardInvocation, handleQueryWithUnboundKey, handleTakeValueForUnboundKey, hash, indexOfAccessibilityElement, isAccessibilityElement, isAccessibilityElementBlock, isEqual, isKindOfClass, isMemberOfClass, isProxy, methodForSelector, methodSignatureForSelector, mutableArrayValueForKey, mutableArrayValueForKeyPath, mutableCopy, mutableOrderedSetValueForKey, mutableOrderedSetValueForKeyPath, mutableSetValueForKey, mutableSetValueForKeyPath, observationInfo, observeValueForKeyPathOfObjectChangeContext, performSelector, performSelectorInBackgroundWithObject, performSelectorOnMainThreadWithObjectWaitUntilDone, performSelectorOnMainThreadWithObjectWaitUntilDoneModes, performSelectorOnThreadWithObjectWaitUntilDone, performSelectorOnThreadWithObjectWaitUntilDoneModes, performSelectorWithObject, performSelectorWithObjectAfterDelay, performSelectorWithObjectAfterDelayInModes, performSelectorWithObjectWithObject, prepareForInterfaceBuilder, provideImageDataBytesPerRowOrigin_Size_UserInfo, removeObserverForKeyPath, removeObserverForKeyPathContext, replacementObjectForCoder, replacementObjectForKeyedArchiver, respondsToSelector, self, setAccessibilityActivateBlock, setAccessibilityActivationPoint, setAccessibilityActivationPointBlock, setAccessibilityAttributedHint, setAccessibilityAttributedHintBlock, setAccessibilityAttributedLabel, setAccessibilityAttributedLabelBlock, setAccessibilityAttributedUserInputLabels, setAccessibilityAttributedUserInputLabelsBlock, setAccessibilityAttributedValue, setAccessibilityAttributedValueBlock, setAccessibilityContainerType, setAccessibilityContainerTypeBlock, setAccessibilityCustomActions, setAccessibilityCustomActionsBlock, setAccessibilityCustomRotors, setAccessibilityCustomRotorsBlock, setAccessibilityDecrementBlock, setAccessibilityDirectTouchOptions, setAccessibilityDragSourceDescriptors, setAccessibilityDropPointDescriptors, setAccessibilityElements, setAccessibilityElementsBlock, setAccessibilityElementsHidden, setAccessibilityElementsHiddenBlock, setAccessibilityFrame, setAccessibilityFrameBlock, setAccessibilityHeaderElementsBlock, setAccessibilityHint, setAccessibilityHintBlock, setAccessibilityIdentifierBlock, setAccessibilityIncrementBlock, setAccessibilityLabel, setAccessibilityLabelBlock, setAccessibilityLanguage, setAccessibilityLanguageBlock, setAccessibilityMagicTapBlock, setAccessibilityNavigationStyle, setAccessibilityNavigationStyleBlock, setAccessibilityPath, setAccessibilityPathBlock, setAccessibilityPerformEscapeBlock, setAccessibilityRespondsToUserInteraction, setAccessibilityRespondsToUserInteractionBlock, setAccessibilityShouldGroupAccessibilityChildrenBlock, setAccessibilityTextualContext, setAccessibilityTextualContextBlock, setAccessibilityTraits, setAccessibilityTraitsBlock, setAccessibilityUserInputLabels, setAccessibilityUserInputLabelsBlock, setAccessibilityValue, setAccessibilityValueBlock, setAccessibilityViewIsModal, setAccessibilityViewIsModalBlock, setAutomationElements, setAutomationElementsBlock, setIsAccessibilityElement, setIsAccessibilityElementBlock, setNilValueForKey, setObservationInfo, setShouldGroupAccessibilityChildren, setValueForKey, setValueForKeyPath, setValueForUndefinedKey, setValuesForKeysWithDictionary, shouldGroupAccessibilityChildren, storedValueForKey, superclass, takeStoredValueForKey, takeValueForKey, takeValueForKeyPath, takeValuesFromDictionary, unableToSetNilForKey, validateValueForKeyError, validateValueForKeyPathError, valueForKey, valueForKeyPath, valueForUndefinedKey, valuesForKeys, willChangeValueForKey, willChangeValueForKeyWithSetMutationUsingObjects, willChangeValuesAtIndexesForKey

Methods inherited from class org.moe.natj.objc.ObjCObject

equals, toString

Methods inherited from class org.moe.natj.general.NativeObject

finalize, getPeer, getPeerPointer

Methods inherited from class java.lang.Object

clone, getClass, hashCode, notify, notifyAll, wait, wait, wait

Constructor Detail

MPSNNOptimizerStochasticGradientDescent

protected MPSNNOptimizerStochasticGradientDescent(org.moe.natj.general.Pointer peer)

Method Detail

accessInstanceVariablesDirectly

public static boolean accessInstanceVariablesDirectly()

alloc

public static MPSNNOptimizerStochasticGradientDescent alloc()

allocWithZone

public static MPSNNOptimizerStochasticGradientDescent allocWithZone(org.moe.natj.general.ptr.VoidPtr zone)

automaticallyNotifiesObserversForKey

public static boolean automaticallyNotifiesObserversForKey(@NotNull
                                                           @NotNull java.lang.String key)

cancelPreviousPerformRequestsWithTarget

public static void cancelPreviousPerformRequestsWithTarget(@NotNull
                                                           @NotNull java.lang.Object aTarget)

cancelPreviousPerformRequestsWithTargetSelectorObject

public static void cancelPreviousPerformRequestsWithTargetSelectorObject(@NotNull
                                                                         @NotNull java.lang.Object aTarget,
                                                                         @NotNull
                                                                         @NotNull org.moe.natj.objc.SEL aSelector,
                                                                         @Nullable
                                                                         @Nullable java.lang.Object anArgument)

classFallbacksForKeyedArchiver

@NotNull
public static @NotNull NSArray<java.lang.String> classFallbacksForKeyedArchiver()

classForKeyedUnarchiver

@NotNull
public static @NotNull org.moe.natj.objc.Class classForKeyedUnarchiver()

debugDescription_static

public static java.lang.String debugDescription_static()

description_static

public static java.lang.String description_static()

encodeToCommandBufferBatchNormalizationGradientStateBatchNormalizationSourceStateInputMomentumVectorsResultState

public void encodeToCommandBufferBatchNormalizationGradientStateBatchNormalizationSourceStateInputMomentumVectorsResultState(@NotNull
                                                                                                                             @NotNull MTLCommandBuffer commandBuffer,
                                                                                                                             @NotNull
                                                                                                                             @NotNull MPSCNNBatchNormalizationState batchNormalizationGradientState,
                                                                                                                             @NotNull
                                                                                                                             @NotNull MPSCNNBatchNormalizationState batchNormalizationSourceState,
                                                                                                                             @Nullable
                                                                                                                             @Nullable NSArray<? extends MPSVector> inputMomentumVectors,
                                                                                                                             @NotNull
                                                                                                                             @NotNull MPSCNNNormalizationGammaAndBetaState resultState)

Encode an MPSNNOptimizerStochasticGradientDescent object to a command buffer to perform out of place update The following operations would be applied useNesterov == NO: m[t] = momentumScale * m[t-1] + learningRate * g variable = variable - m[t] useNesterov == YES: m[t] = momentumScale * m[t-1] + g variable = variable - (learningRate * (g + m[t] * momentumScale)) inputMomentumVector == nil variable = variable - (learningRate * g) where, g is gradient of error wrt variable m[t] is momentum of gradients it is a state we keep updating every update iteration

Parameters:

commandBuffer - A valid MTLCommandBuffer to receive the encoded kernel.

batchNormalizationGradientState - A valid MPSCNNBatchNormalizationState object which specifies the input state with gradients for this update.

batchNormalizationSourceState - A valid MPSCNNBatchNormalizationState object which specifies the input state with original gamma/beta for this update.

inputMomentumVectors - An array MPSVector object which specifies the gradient momentum vectors which will be updated and overwritten. The index 0 corresponds to gamma, index 1 corresponds to beta, array can be of size 1 in which case beta won't be updated

resultState - A valid MPSCNNNormalizationGammaAndBetaState object which specifies the resultValues state which will be updated and overwritten.

encodeToCommandBufferBatchNormalizationStateInputMomentumVectorsResultState

public void encodeToCommandBufferBatchNormalizationStateInputMomentumVectorsResultState(@NotNull
                                                                                        @NotNull MTLCommandBuffer commandBuffer,
                                                                                        @NotNull
                                                                                        @NotNull MPSCNNBatchNormalizationState batchNormalizationState,
                                                                                        @Nullable
                                                                                        @Nullable NSArray<? extends MPSVector> inputMomentumVectors,
                                                                                        @NotNull
                                                                                        @NotNull MPSCNNNormalizationGammaAndBetaState resultState)

Encode an MPSNNOptimizerStochasticGradientDescent object to a command buffer to perform out of place update The following operations would be applied useNesterov == NO: m[t] = momentumScale * m[t-1] + learningRate * g variable = variable - m[t] useNesterov == YES: m[t] = momentumScale * m[t-1] + g variable = variable - (learningRate * (g + m[t] * momentumScale)) inputMomentumVector == nil variable = variable - (learningRate * g) where, g is gradient of error wrt variable m[t] is momentum of gradients it is a state we keep updating every update iteration

Parameters:

commandBuffer - A valid MTLCommandBuffer to receive the encoded kernel.

batchNormalizationState - A valid MPSCNNBatchNormalizationState object which specifies the input state with gradients and original gamma/beta for this update.

inputMomentumVectors - An array MPSVector object which specifies the gradient momentum vectors which will be updated and overwritten. The index 0 corresponds to gamma, index 1 corresponds to beta, array can be of size 1 in which case beta won't be updated

resultState - A valid MPSCNNNormalizationGammaAndBetaState object which specifies the resultValues state which will be updated and overwritten.

encodeToCommandBufferConvolutionGradientStateConvolutionSourceStateInputMomentumVectorsResultState

public void encodeToCommandBufferConvolutionGradientStateConvolutionSourceStateInputMomentumVectorsResultState(@NotNull
                                                                                                               @NotNull MTLCommandBuffer commandBuffer,
                                                                                                               @NotNull
                                                                                                               @NotNull MPSCNNConvolutionGradientState convolutionGradientState,
                                                                                                               @NotNull
                                                                                                               @NotNull MPSCNNConvolutionWeightsAndBiasesState convolutionSourceState,
                                                                                                               @Nullable
                                                                                                               @Nullable NSArray<? extends MPSVector> inputMomentumVectors,
                                                                                                               @NotNull
                                                                                                               @NotNull MPSCNNConvolutionWeightsAndBiasesState resultState)

Encode an MPSNNOptimizerStochasticGradientDescent object to a command buffer to perform out of place update The following operations would be applied useNesterov == NO: m[t] = momentumScale * m[t-1] + learningRate * g variable = variable - m[t] useNesterov == YES: m[t] = momentumScale * m[t-1] + g variable = variable - (learningRate * (g + m[t] * momentumScale)) inputMomentumVector == nil variable = variable - (learningRate * g) where, g is gradient of error wrt variable m[t] is momentum of gradients it is a state we keep updating every update iteration

Parameters:

commandBuffer - A valid MTLCommandBuffer to receive the encoded kernel.

convolutionGradientState - A valid MPSCNNConvolutionGradientState object which specifies the input state with gradients for this update.

convolutionSourceState - A valid MPSCNNConvolutionWeightsAndBiasesState object which specifies the input state with values to be updated.

inputMomentumVectors - An array MPSVector object which specifies the gradient momentum vectors which will be updated and overwritten. The index 0 corresponds to weights, index 1 corresponds to biases, array can be of size 1 in which case biases won't be updated

resultState - A valid MPSCNNConvolutionWeightsAndBiasesState object which specifies the resultValues state which will be updated and overwritten.

encodeToCommandBufferInputGradientMatrixInputValuesMatrixInputMomentumMatrixResultValuesMatrix

public void encodeToCommandBufferInputGradientMatrixInputValuesMatrixInputMomentumMatrixResultValuesMatrix(@NotNull
                                                                                                           @NotNull MTLCommandBuffer commandBuffer,
                                                                                                           @NotNull
                                                                                                           @NotNull MPSMatrix inputGradientMatrix,
                                                                                                           @NotNull
                                                                                                           @NotNull MPSMatrix inputValuesMatrix,
                                                                                                           @Nullable
                                                                                                           @Nullable MPSMatrix inputMomentumMatrix,
                                                                                                           @NotNull
                                                                                                           @NotNull MPSMatrix resultValuesMatrix)

API-Since: 13.0

encodeToCommandBufferInputGradientVectorInputValuesVectorInputMomentumVectorResultValuesVector

public void encodeToCommandBufferInputGradientVectorInputValuesVectorInputMomentumVectorResultValuesVector(@NotNull
                                                                                                           @NotNull MTLCommandBuffer commandBuffer,
                                                                                                           @NotNull
                                                                                                           @NotNull MPSVector inputGradientVector,
                                                                                                           @NotNull
                                                                                                           @NotNull MPSVector inputValuesVector,
                                                                                                           @Nullable
                                                                                                           @Nullable MPSVector inputMomentumVector,
                                                                                                           @NotNull
                                                                                                           @NotNull MPSVector resultValuesVector)

Encode an MPSNNOptimizerStochasticGradientDescent object to a command buffer to perform out of place update The following operations would be applied useNesterov == NO: m[t] = momentumScale * m[t-1] + learningRate * g variable = variable - m[t] useNesterov == YES: m[t] = momentumScale * m[t-1] + g variable = variable - (learningRate * (g + m[t] * momentumScale)) inputMomentumVector == nil variable = variable - (learningRate * g) where, g is gradient of error wrt variable m[t] is momentum of gradients it is a state we keep updating every update iteration

Parameters:

commandBuffer - A valid MTLCommandBuffer to receive the encoded kernel.

inputGradientVector - A valid MPSVector object which specifies the input vector of gradients for this update.

inputValuesVector - A valid MPSVector object which specifies the input vector of values to be updated.

inputMomentumVector - A valid MPSVector object which specifies the gradient momentum vector which will be updated and overwritten.

resultValuesVector - A valid MPSVector object which specifies the resultValues vector which will be updated and overwritten.

hash_static

public static long hash_static()

init

public MPSNNOptimizerStochasticGradientDescent init()

Overrides:

init in class MPSNNOptimizer

initWithCoder

public MPSNNOptimizerStochasticGradientDescent initWithCoder(@NotNull
                                                             @NotNull NSCoder aDecoder)

Description copied from interface: NSCoding

NS_DESIGNATED_INITIALIZER

Specified by:

initWithCoder in interface NSCoding

Overrides:

initWithCoder in class MPSNNOptimizer

initWithCoderDevice

public MPSNNOptimizerStochasticGradientDescent initWithCoderDevice(@NotNull
                                                                   @NotNull NSCoder aDecoder,
                                                                   @NotNull
                                                                   @NotNull java.lang.Object device)

Description copied from class: MPSKernel

NSSecureCoding compatability While the standard NSSecureCoding/NSCoding method -initWithCoder: should work, since the file can't know which device your data is allocated on, we have to guess and may guess incorrectly. To avoid that problem, use initWithCoder:device instead.

Overrides:

initWithCoderDevice in class MPSNNOptimizer

Parameters:

aDecoder - The NSCoder subclass with your serialized MPSKernel

device - The MTLDevice on which to make the MPSKernel

Returns:

A new MPSKernel object, or nil if failure. API-Since: 11.0

initWithDevice

public MPSNNOptimizerStochasticGradientDescent initWithDevice(@NotNull
                                                              @NotNull java.lang.Object device)

Description copied from class: MPSKernel

Standard init with default properties per filter type

Overrides:

initWithDevice in class MPSNNOptimizer

Parameters:

device - The device that the filter will be used on. May not be NULL.

Returns:

a pointer to the newly initialized object. This will fail, returning nil if the device is not supported. Devices must be MTLFeatureSet_iOS_GPUFamily2_v1 or later.

initWithDeviceLearningRate

public MPSNNOptimizerStochasticGradientDescent initWithDeviceLearningRate(@NotNull
                                                                          @NotNull MTLDevice device,
                                                                          float learningRate)

Convenience initialization for the momentum update

Parameters:

device - The device on which the kernel will execute.

learningRate - The learningRate which will be applied

Returns:

A valid MPSNNOptimizerStochasticGradientDescent object or nil, if failure.

initWithDeviceMomentumScaleUseNestrovMomentumOptimizerDescriptor

public MPSNNOptimizerStochasticGradientDescent initWithDeviceMomentumScaleUseNestrovMomentumOptimizerDescriptor(@NotNull
                                                                                                                @NotNull MTLDevice device,
                                                                                                                float momentumScale,
                                                                                                                boolean useNestrovMomentum,
                                                                                                                @NotNull
                                                                                                                @NotNull MPSNNOptimizerDescriptor optimizerDescriptor)

instanceMethodForSelector

public static NSObject.Function_instanceMethodForSelector_ret instanceMethodForSelector(org.moe.natj.objc.SEL aSelector)

instanceMethodSignatureForSelector

public static NSMethodSignature instanceMethodSignatureForSelector(org.moe.natj.objc.SEL aSelector)

instancesRespondToSelector

public static boolean instancesRespondToSelector(org.moe.natj.objc.SEL aSelector)

isSubclassOfClass

public static boolean isSubclassOfClass(org.moe.natj.objc.Class aClass)

keyPathsForValuesAffectingValueForKey

@NotNull
public static @NotNull NSSet<java.lang.String> keyPathsForValuesAffectingValueForKey(@NotNull
                                                                                     @NotNull java.lang.String key)

momentumScale

public float momentumScale()

[@property] momentumScale The momentumScale at which we update momentum for values array Default value is 0.0

new_objc

public static MPSNNOptimizerStochasticGradientDescent new_objc()

resolveClassMethod

public static boolean resolveClassMethod(org.moe.natj.objc.SEL sel)

resolveInstanceMethod

public static boolean resolveInstanceMethod(org.moe.natj.objc.SEL sel)

setVersion_static

public static void setVersion_static(long aVersion)

superclass_static

public static org.moe.natj.objc.Class superclass_static()

supportsSecureCoding

public static boolean supportsSecureCoding()

_supportsSecureCoding

public boolean _supportsSecureCoding()

Description copied from interface: NSSecureCoding

This property must return YES on all classes that allow secure coding. Subclasses of classes that adopt NSSecureCoding and override initWithCoder: must also override this method and return YES. The Secure Coding Guide should be consulted when writing methods that decode data.

Specified by:

_supportsSecureCoding in interface NSSecureCoding

Overrides:

_supportsSecureCoding in class MPSNNOptimizer

useNestrovMomentum

public boolean useNestrovMomentum()

version_static

public static long version_static()

initWithDeviceMomentumScaleUseNesterovMomentumOptimizerDescriptor

public MPSNNOptimizerStochasticGradientDescent initWithDeviceMomentumScaleUseNesterovMomentumOptimizerDescriptor(@NotNull
                                                                                                                 @NotNull MTLDevice device,
                                                                                                                 float momentumScale,
                                                                                                                 boolean useNesterovMomentum,
                                                                                                                 @NotNull
                                                                                                                 @NotNull MPSNNOptimizerDescriptor optimizerDescriptor)

Full initialization for the momentum update

Parameters:

device - The device on which the kernel will execute.

momentumScale - The momentumScale to update momentum for values array

useNesterovMomentum - Use the Nesterov style momentum update

optimizerDescriptor - The optimizerDescriptor which will have a bunch of properties to be applied

Returns:

A valid MPSNNOptimizerMomentum object or nil, if failure. API-Since: 14.0

useNesterovMomentum

public boolean useNesterovMomentum()

[@property] useNesterovMomentum Nesterov momentum is considered an improvement on the usual momentum update Default value is NO [@note] Maps to old useNestrovMomentum property API-Since: 14.0

useStoredAccessor

@Deprecated
public static boolean useStoredAccessor()

Deprecated.