ResourceSparseApplyCenteredRmsProp

public final class ResourceSparseApplyCenteredRmsProp

Update '*var' according to the centered RMSProp algorithm.

The centered RMSProp algorithm uses an estimate of the centered second moment (i.e., the variance) for normalization, as opposed to regular RMSProp, which uses the (uncentered) second moment. This often helps with training, but is slightly more expensive in terms of computation and memory.

Note that in dense implementation of this algorithm, mg, ms, and mom will update even if the grad is zero, but in this sparse implementation, mg, ms, and mom will not update in iterations during which the grad is zero.

mean_square = decay * mean_square + (1-decay) * gradient ** 2 mean_grad = decay * mean_grad + (1-decay) * gradient Delta = learning_rate * gradient / sqrt(mean_square + epsilon - mean_grad ** 2)

ms <- rho * ms_{t-1} + (1-rho) * grad * grad mom <- momentum * mom_{t-1} + lr * grad / sqrt(ms + epsilon) var <- var - mom

Nested Classes

class

ResourceSparseApplyCenteredRmsProp.Options

Optional attributes for ResourceSparseApplyCenteredRmsProp

Constants

String

OP_NAME

The name of this op, as known by TensorFlow core engine

Public Methods

static <T extends TType> ResourceSparseApplyCenteredRmsProp	create(Scope scope, Operand<?> var, Operand<?> mg, Operand<?> ms, Operand<?> mom, Operand<T> lr, Operand<T> rho, Operand<T> momentum, Operand<T> epsilon, Operand<T> grad, Operand<? extends TNumber> indices, Options... options) Factory method to create a class wrapping a new ResourceSparseApplyCenteredRmsProp operation.
static ResourceSparseApplyCenteredRmsProp.Options	useLocking(Boolean useLocking)

Inherited Methods

From class org.tensorflow.op.RawOp

final boolean	equals(Object obj)
final int	hashCode()
Operation	op() Return this unit of computation as a single Operation.
final String	toString()

From class java.lang.Object

boolean	equals(Object arg0)
final Class<?>	getClass()
int	hashCode()
final void	notify()
final void	notifyAll()
String	toString()
final void	wait(long arg0, int arg1)
final void	wait(long arg0)
final void	wait()

From interface org.tensorflow.op.Op

abstract ExecutionEnvironment	env() Return the execution environment this op was created in.
abstract Operation	op() Return this unit of computation as a single Operation.

Constants

public static final String OP_NAME

The name of this op, as known by TensorFlow core engine

Constant Value: "ResourceSparseApplyCenteredRMSProp"

Public Methods

public static ResourceSparseApplyCenteredRmsProp create (Scope scope, Operand<?> var, Operand<?> mg, Operand<?> ms, Operand<?> mom, Operand<T> lr, Operand<T> rho, Operand<T> momentum, Operand<T> epsilon, Operand<T> grad, Operand<? extends TNumber> indices, Options... options)

Factory method to create a class wrapping a new ResourceSparseApplyCenteredRmsProp operation.

Parameters

scope	current scope
var	Should be from a Variable().
mg	Should be from a Variable().
ms	Should be from a Variable().
mom	Should be from a Variable().
lr	Scaling factor. Must be a scalar.
rho	Decay rate. Must be a scalar.
epsilon	Ridge term. Must be a scalar.
grad	The gradient.
indices	A vector of indices into the first dimension of var, ms and mom.
options	carries optional attributes values

Returns

• a new instance of ResourceSparseApplyCenteredRmsProp

public static ResourceSparseApplyCenteredRmsProp.Options useLocking (Boolean useLocking)

Parameters

useLocking	If `True`, updating of the var, mg, ms, and mom tensors is protected by a lock; otherwise the behavior is undefined, but may exhibit less contention.