deepboof.backward

Interface DBatchNorm<T extends deepboof.Tensor<T>>

All Superinterfaces:

deepboof.forward.BatchNorm, deepboof.DFunction<T>, deepboof.Function<T>

All Known Subinterfaces:

DFunctionBatchNorm<T>, DSpatialBatchNorm<T>

All Known Implementing Classes:

BaseDBatchNorm_F64, DFunctionBatchNorm_F64, DSpatialBatchNorm_F64

public interface DBatchNorm<T extends deepboof.Tensor<T>>
extends deepboof.forward.BatchNorm, deepboof.DFunction<T>

Implementation of batch normalization for training networks. The mean and standard deviation is always computed on the forwards pass. Unlike the forward only implementation the only parameters (which are optional) are gamma and beta.

Method Summary

Modifier and Type	Method and Description
T	getMean(T output) Returns the most recently computed mean for each variable in the tensor.
T	getVariance(T output) Returns the most recently computed variance for each variable.

Methods inherited from interface deepboof.forward.BatchNorm

getEPS, hasGammaBeta, setEPS

Methods inherited from interface deepboof.DFunction

backwards, evaluating, isLearning, learning

Methods inherited from interface deepboof.Function

forward, getOutputShape, getParameters, getParameterShapes, getTensorType, initialize, setParameters

Method Detail

getMean

T getMean(T output)

Returns the most recently computed mean for each variable in the tensor.

Parameters:

output - Storage for mean tensor. Is reshaped. If null a new instance will be declared

getVariance

T getVariance(T output)

Returns the most recently computed variance for each variable. This will be the actual variance not something that has been adjusted by adding EPS to it.

Parameters:

output - Storage for variance tensor. Is reshaped. If null a new instance will be declared