LossApiShape
Enums · AiDotNet.Enums
Describes the method signature shape a loss function uses for its primary calculation. The test scaffold generator uses this to select the correct test base class.
Fields
| Field | Summary |
|---|---|
LossApiShape ComplexInterleaved |
Standard CalculateLoss(Vector, Vector) but inputs are complex-interleaved pairs [real, imag, real, imag, …]. |
LossApiShape ImageMatrix |
Image-based Calculate(Matrix, Matrix) requiring a feature extractor. |
LossApiShape PairedEmbedding |
Paired embedding CalculateLoss(Vector, Vector, T) with two embedding vectors and a similarity label. |
LossApiShape SelfSupervised |
Self-supervised CreateTask interface, not a standard loss calculation. |
LossApiShape Sequence |
Sequence loss over log-probabilities and label sequences of unknown alignment. |
LossApiShape SparseIndex |
Standard CalculateLoss(Vector, Vector) but predicted and actual have different lengths. |
LossApiShape TargetNoiseMatrix |
Contrastive-style Calculate(Vector, Matrix) with target logits and noise logits. |
LossApiShape TripletMatrix |
Triplet-style CalculateLoss(Matrix, Matrix, Matrix) with anchor, positive, negative. |
LossApiShape VectorVector |
Standard CalculateLoss(Vector, Vector) interface. |
int value__ |