HamiltonianNeuralNetworkOptions

Options & Configuration · AiDotNet.PhysicsInformed.Options

Configuration options for the HamiltonianNeuralNetwork.

Inheritance: ModelOptions → NeuralNetworkOptions → ModelHyperparameterOptions → PhysicsInformedOptions → HamiltonianNeuralNetworkOptions

Constructors

Constructor Summary
HamiltonianNeuralNetworkOptions() Initializes the paper-default HNN options.
HamiltonianNeuralNetworkOptions(HamiltonianNeuralNetworkOptions other) Initializes a value copy of an existing HNN configuration.

Properties

Property Summary
double LearningRate { get; set; } Gets or sets the fixed Adam learning rate.
double SymplecticSolverTolerance { get; set; } Gets or sets the convergence tolerance of the implicit momentum update in Simulate.
double WeightDecay { get; set; } Gets or sets the coupled L2 penalty used by Adam.
int HiddenDimension { get; set; } Gets or sets the width of every hidden layer.
int HiddenLayerCount { get; set; } Gets or sets the number of tanh hidden layers.
int SymplecticSolverMaxIterations { get; set; } Gets or sets the most fixed-point iterations of the implicit momentum update in Simulate.

Methods

Method Summary
void Validate() Validates the paper training recipe and architecture.