crested.tl.metrics.SpearmanCorrelationPerClass#
- class crested.tl.metrics.SpearmanCorrelationPerClass(num_classes, name='multiclass_spearman_correlation', **kwargs)#
Spearman correlation metric for multiclass models.
Attributes table#
Methods table#
Reset the state of the metric. |
|
|
Calculate the result of the metric. |
|
Update the state of the metric. |
Attributes#
Methods#
- SpearmanCorrelationPerClass.reset_state()#
Reset the state of the metric.
- SpearmanCorrelationPerClass.result()#
Calculate the result of the metric.
- SpearmanCorrelationPerClass.update_state(y_true, y_pred, sample_weight=None)#
Update the state of the metric.
- Args:
y_true (tensor): Ground truth labels of shape (batch_size, num_classes). y_pred (tensor): Predicted outputs of shape (batch_size, num_classes).