Table of Contents

Class Trainer4D

Namespace
NeuralNetworks.Trainers
Assembly
NeuralNetworks.dll
public class Trainer4D : Trainer<float[,,,], float[,]>
Inheritance
Trainer<float[,,,], float[,]>
Trainer4D
Inherited Members

Constructors

Trainer4D(Model<float[,,,], float[,]>, Optimizer, SeededRandom, ILogger<Trainer4D>, bool)

public Trainer4D(Model<float[,,,], float[,]> model, Optimizer optimizer, SeededRandom random, ILogger<Trainer4D> logger, bool operationBackendTimingEnabled = false)

Parameters

model Model<float[,,,], float[,]>
optimizer Optimizer
random SeededRandom
logger ILogger<Trainer4D>
operationBackendTimingEnabled bool

Methods

GenerateBatches(float[,,,], float[,], int)

Generates batches of input and output matrices.

protected override IEnumerable<(float[,,,] xBatch, float[,] yBatch)> GenerateBatches(float[,,,] x, float[,] y, int batchSize)

Parameters

x float[,,,]

The input matrix.

y float[,]

The output matrix.

batchSize int

The batch size.

Returns

IEnumerable<(float[,,,] xBatch, float[,] yBatch)>

An enumerable of batches.

GetRows(float[,,,])

protected override float GetRows(float[,,,] x)

Parameters

x float[,,,]

Returns

float

PermuteData(float[,,,], float[,], Random)

protected override void PermuteData(float[,,,] x, float[,] y, Random random)

Parameters

x float[,,,]
y float[,]
random Random