Class MultiScaleTrainingOptions<T>

Configuration options for multi-scale PINN training.

public class MultiScaleTrainingOptions<T>

Inherited Members: object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Properties

Coupling loss weight (balances scale coupling vs individual scale losses).

public T? CouplingWeight { get; set; }

Individual weights for each scale (overrides automatic weighting).

public T[]? ManualScaleWeights { get; set; }

Number of epochs to pre-train each scale before adding the next.

public int ScalePretrainingEpochs { get; set; }

Whether to use adaptive scale weighting during training.

public bool UseAdaptiveScaleWeighting { get; set; }

Whether to train scales sequentially (coarse to fine) or simultaneously.

public bool UseSequentialScaleTraining { get; set; }

Sequential training can be more stable:

First train coarse scale until convergence
Then add fine scale and continue training This is called "progressive training" or "curriculum learning".