public abstract class BaseDBatchNorm_F64 extends BaseDFunction<deepboof.tensors.Tensor_F64> implements DBatchNorm<deepboof.tensors.Tensor_F64>
Tensor_F64.| Modifier and Type | Field and Description |
|---|---|
protected int |
D |
protected double |
EPS |
protected deepboof.tensors.Tensor_F64 |
params |
protected boolean |
requiresGammaBeta |
protected int[] |
shapeVariables |
protected deepboof.tensors.Tensor_F64 |
tensorDiffX |
protected deepboof.tensors.Tensor_F64 |
tensorDMean |
protected deepboof.tensors.Tensor_F64 |
tensorDVar |
protected deepboof.tensors.Tensor_F64 |
tensorDXhat |
protected deepboof.tensors.Tensor_F64 |
tensorMean |
protected deepboof.tensors.Tensor_F64 |
tensorStd |
protected deepboof.tensors.Tensor_F64 |
tensorTmp |
protected deepboof.tensors.Tensor_F64 |
tensorXhat |
learningMode| Constructor and Description |
|---|
BaseDBatchNorm_F64(boolean requiresGammaBeta) |
| Modifier and Type | Method and Description |
|---|---|
void |
_initialize() |
void |
_setParameters(java.util.List<deepboof.tensors.Tensor_F64> parameters) |
protected abstract int[] |
createShapeVariables(int[] shapeInput)
Create the shape for all the variables which are being normalized
|
double |
getEPS() |
deepboof.tensors.Tensor_F64 |
getMean(deepboof.tensors.Tensor_F64 output)
Returns the most recently computed mean for each variable in the tensor.
|
java.lang.Class<deepboof.tensors.Tensor_F64> |
getTensorType() |
deepboof.tensors.Tensor_F64 |
getVariance(deepboof.tensors.Tensor_F64 output)
Returns the most recently computed variance for each variable.
|
boolean |
hasGammaBeta() |
void |
setEPS(double EPS) |
_backwards, backwards, evaluating, isLearning, learning_forward, forward, getOutputShape, getParameters, getParameterShapes, initialize, setParametersprotected boolean requiresGammaBeta
protected deepboof.tensors.Tensor_F64 tensorMean
protected deepboof.tensors.Tensor_F64 tensorStd
protected deepboof.tensors.Tensor_F64 tensorXhat
protected deepboof.tensors.Tensor_F64 tensorDVar
protected deepboof.tensors.Tensor_F64 tensorDMean
protected deepboof.tensors.Tensor_F64 tensorDXhat
protected deepboof.tensors.Tensor_F64 tensorDiffX
protected deepboof.tensors.Tensor_F64 tensorTmp
protected int[] shapeVariables
protected int D
protected deepboof.tensors.Tensor_F64 params
protected double EPS
public void _initialize()
_initialize in class deepboof.impl.forward.standard.BaseFunction<deepboof.tensors.Tensor_F64>protected abstract int[] createShapeVariables(int[] shapeInput)
shapeInput - Shape of input tensor (without mini-batch)public void _setParameters(java.util.List<deepboof.tensors.Tensor_F64> parameters)
_setParameters in class deepboof.impl.forward.standard.BaseFunction<deepboof.tensors.Tensor_F64>public double getEPS()
getEPS in interface deepboof.forward.BatchNormpublic void setEPS(double EPS)
setEPS in interface deepboof.forward.BatchNormpublic boolean hasGammaBeta()
hasGammaBeta in interface deepboof.forward.BatchNormpublic java.lang.Class<deepboof.tensors.Tensor_F64> getTensorType()
getTensorType in interface deepboof.Function<deepboof.tensors.Tensor_F64>public deepboof.tensors.Tensor_F64 getMean(deepboof.tensors.Tensor_F64 output)
DBatchNormgetMean in interface DBatchNorm<deepboof.tensors.Tensor_F64>output - Storage for mean tensor. Is reshaped. If null a new instance will be declaredpublic deepboof.tensors.Tensor_F64 getVariance(deepboof.tensors.Tensor_F64 output)
DBatchNormgetVariance in interface DBatchNorm<deepboof.tensors.Tensor_F64>output - Storage for variance tensor. Is reshaped. If null a new instance will be declared