@Operator(group="train") public final class ResourceApplyRmsProp extends PrimitiveOp
Note that in dense implementation of this algorithm, ms and mom will update even if the grad is zero, but in this sparse implementation, ms and mom will not update in iterations during which the grad is zero.
mean_square = decay * mean_square + (1-decay) * gradient ** 2 Delta = learning_rate * gradient / sqrt(mean_square + epsilon)
ms <- rho * ms_{t-1} + (1-rho) * grad * grad mom <- momentum * mom_{t-1} + lr * grad / sqrt(ms + epsilon) var <- var - mom
Modifier and Type | Class and Description |
---|---|
static class |
ResourceApplyRmsProp.Options
Optional attributes for
ResourceApplyRmsProp |
operation
Modifier and Type | Method and Description |
---|---|
static <T> ResourceApplyRmsProp |
create(Scope scope,
Operand<?> var,
Operand<?> ms,
Operand<?> mom,
Operand<T> lr,
Operand<T> rho,
Operand<T> momentum,
Operand<T> epsilon,
Operand<T> grad,
ResourceApplyRmsProp.Options... options)
Factory method to create a class wrapping a new ResourceApplyRmsProp operation.
|
static ResourceApplyRmsProp.Options |
useLocking(Boolean useLocking) |
equals, hashCode, op, toString
public static <T> ResourceApplyRmsProp create(Scope scope, Operand<?> var, Operand<?> ms, Operand<?> mom, Operand<T> lr, Operand<T> rho, Operand<T> momentum, Operand<T> epsilon, Operand<T> grad, ResourceApplyRmsProp.Options... options)
scope
- current scopevar
- 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.momentum
- epsilon
- Ridge term. Must be a scalar.grad
- The gradient.options
- carries optional attributes valuespublic static ResourceApplyRmsProp.Options useLocking(Boolean useLocking)
useLocking
- If `True`, updating of the var, ms, and mom tensors is protected
by a lock; otherwise the behavior is undefined, but may exhibit less
contention.Copyright © 2022. All rights reserved.