IMultiFidelityTrainingHistory<T>
Interfaces · AiDotNet.PhysicsInformed.Interfaces
Extended training history interface for multi-fidelity PINN training.
Known implementations: MultiFidelityTrainingHistory<T>
How It Works
For Beginners: Multi-fidelity training uses data from multiple sources with different accuracy levels:
- Low-fidelity: Cheap but less accurate (e.g., coarse simulations, simplified models)
- High-fidelity: Expensive but accurate (e.g., fine simulations, experiments)
This interface tracks metrics for each fidelity level during training.
Properties
| Property | Summary |
|---|---|
List<T> CorrelationLosses { get; } |
Gets the correlation losses per epoch (measures agreement between fidelity levels). |
List<T> HighFidelityLosses { get; } |
Gets the high-fidelity data losses per epoch. |
List<T> Losses { get; } |
Gets the total losses per epoch (combined from all fidelity levels). |
List<T> LowFidelityLosses { get; } |
Gets the low-fidelity data losses per epoch. |
List<T> PhysicsLosses { get; } |
Gets the PDE residual losses per epoch. |
Methods
| Method | Summary |
|---|---|
void AddEpoch(T totalLoss, T lowFidelityLoss, T highFidelityLoss, T correlationLoss, T physicsLoss) |
Records metrics for a training epoch. |