Module: tfma.post_export_metrics

Library containing helpers for adding post export metrics for evaluation.

These post export metrics can be included in the add_post_export_metrics parameter of Evaluate to compute them.

Functions

auc(...): This is the function that the user calls.

auc_plots(...): This is the function that the user calls.

calibration(...): This is the function that the user calls.

calibration_plot_and_prediction_histogram(...): This is the function that the user calls.

confusion_matrix_at_thresholds(...): This is the function that the user calls.

example_count(...): This is the function that the user calls.

example_weight(...): This is the function that the user calls.

fairness_auc(...): This is the function that the user calls.

fairness_indicators(...): This is the function that the user calls.

mean_absolute_error(...): This is the function that the user calls.

mean_squared_error(...): This is the function that the user calls.

precision_at_k(...): This is the function that the user calls.

recall_at_k(...): This is the function that the user calls.

root_mean_squared_error(...): This is the function that the user calls.

squared_pearson_correlation(...): This is the function that the user calls.

DEFAULT_KEY_PREFERENCE ('logistic', 'predictions', 'probabilities', 'logits')