LossTestInputFormat

Enums · AiDotNet.Enums

Describes what kind of test data a loss function expects. The test scaffold generator uses this to create appropriate test vectors.

Fields

Field Summary
LossTestInputFormat Continuous Standard continuous values in [0, 1] range.
LossTestInputFormat CriticScores Wasserstein critic scores with signed labels.
LossTestInputFormat MarginBased Margin-based with predictions in [0, 1] and binary labels.
LossTestInputFormat OrdinalCategories Ordinal category indices.
LossTestInputFormat ProbabilityDistribution Probability distribution where values are in [0, 1].
LossTestInputFormat RawLogits Raw logits (unbounded real values) with one-hot or soft targets.
LossTestInputFormat SegmentationMask Segmentation masks in [0, 1] range.
LossTestInputFormat SignedLabels Signed labels in {-1, +1} with predictions as real-valued scores.
LossTestInputFormat SimilarityLabels Binary similarity labels in {0, 1} with distance-based predictions.
int value__