Package | Description |
---|---|
org.nd4j.linalg.api.ops | |
org.nd4j.linalg.api.ops.factory | |
org.nd4j.linalg.api.ops.impl.transforms | |
org.nd4j.linalg.api.ops.impl.transforms.gradient |
Modifier and Type | Class and Description |
---|---|
class |
BaseGradientOp
A gradient op always makes the following assumptions:
there is always a y (beacuse of backpropagating
or using the chain rule)
and that it is special exec (for now)
This op opType sis meant to be used
to build derivative operations.
|
Modifier and Type | Method and Description |
---|---|
GradientOp |
OpFactory.createGradientOp(String name,
INDArray x,
INDArray y,
INDArray z) |
GradientOp |
DefaultOpFactory.createGradientOp(String name,
INDArray x,
INDArray y,
INDArray z) |
Modifier and Type | Class and Description |
---|---|
class |
LogSigmoidDerivative
LogSigmoid derivative
|
Modifier and Type | Class and Description |
---|---|
class |
GradientBackwardsMarker |
class |
LogSoftMaxDerivative |
class |
SigmoidDerivative |
class |
SoftMaxDerivative |
class |
TanhDerivative |
Copyright © 2018. All rights reserved.