keras-team/keras · error · ValueError

Invalid AUC summation method value: "{key}". Expected values

Error message

Invalid AUC summation method value: "{key}". Expected values are ["interpolation", "majoring", "minoring"]

What it means

AUC's summation_method controls how area under the curve is approximated. AUCSummationMethod.from_str() accepts only 'interpolation', 'majoring' or 'minoring' (case-insensitive) and raises this ValueError otherwise. Strings like 'trapezoid' or 'interpolate' are rejected.

Source

Thrown at keras/src/metrics/metrics_utils.py:92

      summation for decreasing intervals.
    * 'majoring': Applies right summation for increasing intervals and left
      summation for decreasing intervals.
    """

    INTERPOLATION = "interpolation"
    MAJORING = "majoring"
    MINORING = "minoring"

    @staticmethod
    def from_str(key):
        if key in ("interpolation", "Interpolation"):
            return AUCSummationMethod.INTERPOLATION
        elif key in ("majoring", "Majoring"):
            return AUCSummationMethod.MAJORING
        elif key in ("minoring", "Minoring"):
            return AUCSummationMethod.MINORING
        else:
            raise ValueError(
                f'Invalid AUC summation method value: "{key}". '
                'Expected values are ["interpolation", "majoring", "minoring"]'
            )


def _update_confusion_matrix_variables_optimized(
    variables_to_update,
    y_true,
    y_pred,
    thresholds,
    multi_label=False,
    sample_weights=None,
    label_weights=None,
    thresholds_with_epsilon=False,
):
    """Update confusion matrix variables with memory efficient alternative.

    Note that the thresholds need to be evenly distributed within the list, eg,

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Use exactly 'interpolation', 'majoring', or 'minoring'.
2. Keep the default summation_method='interpolation' unless you specifically need Riemann majoring/minoring behavior.

Example fix

# before
auc = keras.metrics.AUC(summation_method='trapezoid')

# after
auc = keras.metrics.AUC(summation_method='interpolation')

VALID_SUMMATION = {'interpolation', 'majoring', 'minoring'}
def check_summation(s):
    if s.lower() not in VALID_SUMMATION:
        raise ValueError(f'summation_method must be one of {sorted(VALID_SUMMATION)}')
    return s

When it happens

Trigger: Calling keras.metrics.AUC(summation_method='trapezoid') or summation_method='interpolate' - anything other than the three accepted keys.

Common situations: Assuming numpy.trapz-style naming ('trapezoid') transfers to Keras; typos in hyperparameter configs.

Understand the failure class

Background: Invalid enum value errors: "Unknown type", "Invalid scope", "must be one of" — when a string is not on the library's allowed list — this error's family across 23 libraries.

Related errors

• Invalid AUC curve value: "{key}". Expected values are ["PR",
• Invalid `curve` argument value "{curve}". Expected one of: {
• Invalid `summation_method` argument value "{summation_method
• Argument `num_thresholds` must be an integer > 1. Received:
• `num_labels` is needed only when `multi_label` is True.

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/0c9cb875732877cb. Report an issue: GitHub.