Class NeuralArchitectureSearch<T>

Namespace

AiDotNet.AutoML

Assembly

AiDotNet.dll

Neural Architecture Search implementation with gradient-based (DARTS) support

public class NeuralArchitectureSearch<T>

Type Parameters

T

The numeric type for calculations

Inheritance

object

NeuralArchitectureSearch<T>

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Constructors

NeuralArchitectureSearch(NeuralArchitectureSearchStrategy, int)

public NeuralArchitectureSearch(NeuralArchitectureSearchStrategy strategy = NeuralArchitectureSearchStrategy.GradientBased, int maxEpochs = 50)

Parameters

strategy NeuralArchitectureSearchStrategy

maxEpochs int

Properties

BestArchitecture

public Architecture<T>? BestArchitecture { get; }

Property Value

Architecture<T>

BestScore

public T BestScore { get; }

Property Value

T

Status

public AutoMLStatus Status { get; }

Property Value

AutoMLStatus

Methods

SearchAsync(Tensor<T>, Tensor<T>, Tensor<T>, Tensor<T>, CancellationToken)

Runs the neural architecture search

public Task<Architecture<T>> SearchAsync(Tensor<T> trainData, Tensor<T> trainLabels, Tensor<T> valData, Tensor<T> valLabels, CancellationToken cancellationToken = default)

Parameters

trainData Tensor<T>

trainLabels Tensor<T>

valData Tensor<T>

valLabels Tensor<T>

cancellationToken CancellationToken

Returns

Task<Architecture<T>>