Crate torsh_metrics 

Comprehensive evaluation metrics for ToRSh

This module provides PyTorch-compatible metrics for model evaluation, built on top of SciRS2’s comprehensive metrics library.

Re-exports

pub use deep_learning::BleuScore;

pub use deep_learning::DeepLearningMetrics;

pub use deep_learning::RougeMetrics;

pub use deep_learning::RougeScore;

pub use deep_learning::RougeType;

pub use deep_learning::SimilarityType;

pub use deep_learning::VectorizedFidScore;

pub use deep_learning::VectorizedInceptionScore;

pub use deep_learning::VectorizedPerplexity;

pub use deep_learning::VectorizedSemanticSimilarity;

pub use classification::ConfusionMatrix;

pub use classification::MultiClassMetrics;

pub use classification::ThresholdMetrics;

pub use ranking::IRMetrics;

pub use uncertainty::BayesianUncertainty;

pub use uncertainty::CalibrationMetrics;

pub use uncertainty::EnsembleUncertainty;

pub use uncertainty::MCDropoutUncertainty;

pub use uncertainty::UncertaintyDecomposition;

pub use fairness::FairnessMetrics;

pub use statistics::BootstrapResult;

pub use statistics::CrossValidationResult;

pub use statistics::HypothesisTestResult;

pub use gpu::GpuAccuracy;

pub use gpu::GpuBatchMetrics;

pub use gpu::GpuConfusionMatrix;

pub use parallel::ParallelAccuracy;

pub use parallel::ParallelConfusionMatrix;

pub use parallel::ParallelMetricCollection;

pub use reporting::ComparisonReport;

pub use reporting::MetricReport;

pub use reporting::ReportBuilder;

pub use reporting::ReportFormat;

pub use memory_efficient::ChunkedEvaluator;

pub use memory_efficient::MemoryEfficientAccuracy;

pub use memory_efficient::MemoryEfficientMAE;

pub use memory_efficient::MemoryEfficientMSE;

pub use memory_efficient::OnlineConfusionMatrix;

pub use memory_efficient::StreamingMetric;

pub use tensorboard::MetricLogger as TensorBoardLogger;

pub use tensorboard::TensorBoardWriter;

pub use mlflow::ExperimentTracker;

pub use mlflow::MLflowClient;

pub use mlflow::MLflowRun;

pub use visualization::CalibrationCurvePlot;

pub use visualization::ConfusionMatrixPlot;

pub use visualization::ExportFormat;

pub use visualization::FeatureImportancePlot;

pub use visualization::InteractiveDashboard;

pub use visualization::LatexReportBuilder;

pub use visualization::LearningCurvePlot;

pub use visualization::MetricComparisonPlot;

pub use visualization::PRCurvePlot;

pub use visualization::ROCCurvePlot;

pub use visualization::VisualizationAggregator;

pub use advanced_ml::ContinualLearningMetrics;

pub use advanced_ml::DomainAdaptationMetrics;

pub use advanced_ml::FewShotMetrics;

pub use advanced_ml::MetaLearningMetrics;

pub use sklearn_compat::SklearnAccuracy;

pub use sklearn_compat::SklearnF1Score;

pub use sklearn_compat::SklearnMeanAbsoluteError;

pub use sklearn_compat::SklearnMeanSquaredError;

pub use sklearn_compat::SklearnMetric;

pub use sklearn_compat::SklearnPrecision;

pub use sklearn_compat::SklearnR2Score;

pub use sklearn_compat::SklearnRecall;

pub use wandb::LogEntry;

pub use wandb::WandbClient;

pub use model_selection::AICc;

pub use model_selection::CVModelComparison;

pub use model_selection::CVModelSelection;

pub use model_selection::CVScoreType;

pub use model_selection::ModelComparisonReport;

pub use model_selection::MultiModelComparison;

pub use model_selection::AIC;

pub use model_selection::BIC;

pub use model_selection::HQIC;

pub use statistical_tests::FiveByTwoCVTest;

pub use statistical_tests::FriedmanTest;

pub use statistical_tests::KruskalWallisTest;

pub use statistical_tests::MannWhitneyTest;

pub use statistical_tests::McNemarTest;

pub use statistical_tests::NemenyiTest;

pub use statistical_tests::PairedTTest;

pub use statistical_tests::WilcoxonTest;

pub use time_series::dtw_distance;

pub use time_series::error_autocorrelation;

pub use time_series::mape;

pub use time_series::mase;

pub use time_series::mean_directional_accuracy;

pub use time_series::msis;

pub use time_series::smape;

pub use time_series::theil_u;

pub use time_series::tracking_signal;

pub use regression_diagnostics::breusch_pagan_test;

pub use regression_diagnostics::calculate_leverage;

pub use regression_diagnostics::condition_number;

pub use regression_diagnostics::cooks_distance;

pub use regression_diagnostics::dffits;

pub use regression_diagnostics::durbin_watson;

pub use regression_diagnostics::variance_inflation_factor;

pub use regression_diagnostics::RegressionDiagnosticReport;

pub use regression_diagnostics::ResidualDiagnostics;

pub use explainability::attribution_agreement;

pub use explainability::counterfactual_validity;

pub use explainability::explanation_completeness;

pub use explainability::explanation_faithfulness;

pub use explainability::feature_importance_stability;

pub use explainability::feature_monotonicity;

pub use explainability::interaction_strength;

pub use explainability::ExplainabilityMetrics;

pub use robustness::adversarial_accuracy;

pub use robustness::attack_success_rate;

pub use robustness::certified_robustness_radius;

pub use robustness::confidence_stability;

pub use robustness::corruption_robustness;

pub use robustness::gradient_stability;

pub use robustness::noise_sensitivity;

pub use robustness::ood_detection_score;

pub use robustness::robustness_accuracy_tradeoff;

pub use robustness::RobustnessReport;

Modules

advanced_ml

Advanced ML Metrics for Meta-Learning, Few-Shot Learning, Domain Adaptation, and Continual Learning

classification

Classification metrics

clustering

Clustering metrics with comprehensive evaluation algorithms

deep_learning

Deep learning specific metrics with high-performance vectorized implementations

explainability

Explainability and interpretability metrics

fairness

Fairness and bias detection metrics

gpu

GPU-accelerated metrics for high-performance evaluation

memory_efficient

Memory-efficient large dataset evaluation

mlflow

MLflow integration for experiment tracking

model_selection

Model selection metrics for choosing optimal models

parallel

Parallel metric computation for scalability

ranking

Ranking and recommendation metrics with comprehensive implementations

regression

Regression metrics

regression_diagnostics

Regression diagnostic metrics and tools

reporting

Automated metric reporting and visualization

robustness

Robustness and reliability metrics

sklearn_compat

Scikit-learn compatibility layer for torsh-metrics

statistical_tests

Advanced statistical hypothesis testing for model comparison

statistics

Statistical validation and bootstrap confidence intervals

streaming

Streaming (online) metrics for efficient large-scale evaluation

tensorboard

TensorBoard integration for metric logging

time_series

Time series forecasting metrics

uncertainty

Uncertainty quantification metrics

utils

Utility functions for metrics

visualization

Metric visualization utilities

wandb

Weights & Biases (W&B) integration for experiment tracking

Structs

MetricCollection

Metric collection for evaluating multiple metrics at once

Traits

Metric

Base trait for all metrics