MetricFrame¶
fairlearn_fhe.metrics.MetricFrame(*, metrics, y_true, y_pred, sensitive_features, sample_params=None, allow_decrypt=False)
¶
Encrypted-aware MetricFrame factory.
Mirrors :class:fairlearn.metrics.MetricFrame for plaintext inputs
(returns the genuine fairlearn object so isinstance against
:class:fairlearn.metrics.MetricFrame works). For encrypted
y_pred returns an :class:EncryptedMetricFrame with the same
accessors used by the fairness wrappers.
This is a function rather than a class so that callers don't have
to think about which concrete type they got back; introspect via
isinstance(result, EncryptedMetricFrame) if you need to
distinguish.