API Reference
Task & Metric Classes
Core classes for defining selective inference tasks and metrics
- SelectiveInferenceTask
-
Wrap a trained
LightningModuleto attach selection results during inference. - SelectiveMetric
-
Evaluate a metric on full, selected, and rejected subsets.
- RiskCoverageMetric
-
Build a risk-coverage curve from scores and per-sample errors.
- RiskCoverage
-
Container for risk-coverage results.
Embedding based Uncertainty Scores
Uncertainty scoring methods for embedding-based approaches
- scores.CosineScore
-
Returns the KNN-distance based on the cosine distance to the nearest samples.
- scores.CosineClassWiseScore
-
Class-wise version of
scores.knn.CosineScore. - scores.EuclideanScore
-
Returns the KNN-distance based on the Euclidean distance to the nearest samples.
- scores.EuclideanClassWiseScore
-
Class-wise version of
scores.knn.EuclideanScore. - scores.MahalanobisScore
-
Returns the Mahalanobis distance to the training samples distribution.
- scores.MahalanobisClassWiseScore
-
Class-wise version of
.scores.knn.MahalanobisScore. - scores.PCAScore
-
Returns uncertainty scores based on PCA reconstruction errors.
- scores.PyODScore
-
Uncertainty scores based on detectors supplied by PyOD.
- scores.utils.TensorPCA
-
Tensor-based PCA.
Logit based Uncertainty Scores
Uncertainty scoring methods based on model logits
- scores.EnergyScore
-
Free energy of the logit distribution.
- scores.EnergyClassWiseScore
-
Class-wise version of
~seapig.scores.logits.EnergyScore. - scores.EntropyScore
-
Predictive entropy (worst label for multilabel).
- scores.EntropyClassWiseScore
-
Class-wise version of
~seapig.scores.logits.EntropyScore. - scores.MarginScore
-
Negative top-two logit margin.
- scores.MarginClassWiseScore
-
Class-wise version of
~seapig.scores.logits.MarginScore. - scores.SoftmaxScore
-
Negative maximum predicted probability.
- scores.SoftmaxClassWiseScore
-
Class-wise version of
~seapig.scores.logits.SoftmaxScore. - scores.MutualInformationScore
-
Mutual information (BALD) uncertainty score for ensembles / MC-dropout.
- scores.MutualInformationClassWiseScore
-
Class-wise version of
~seapig.scores.logits.MutualInformationScore. - scores.PredictiveVarianceScore
-
Variance of predicted probabilities across members.
- scores.PredictiveVarianceClassWiseScore
-
Class-wise version of
~seapig.scores.logits.PredictiveVarianceScore.
Other Uncertainty Scores
Uncertainty scoring methods based on other principles
- scores.RandomScore
-
Returns random uncertainty scores per sample.
Abstract Classes
Abstract base classes
- scores.UncertaintyScore
-
Abstract Base Class for Uncertainty Scores.
- scores.EmbeddingScore
-
Base class for embedding-based uncertainty scores.
- scores.KNNScore
-
Abstract base class for KNN distance-based uncertainty scores.
- scores.LogitScore
-
Base class for logit-based uncertainty scores.
- scores.ClassWiseScore
-
Base class-wise wrapper.
Utilities
Helpers for progress and logging
- utils.track()
-
Wrap an iterable with a progress bar.
- utils.enable()
-
Globally enable progress reporting.
- utils.disable()
-
Globally disable progress reporting.
- utils.set_backend()
-
Select the progress-bar backend.
- utils.get_backend()
-
Return the name of the currently selected backend.
- utils.get_logger()
-
Return a logger scoped to the seapig package.
- utils.configure_logging()
-
Configure the seapig package logger.