sklearn.metrics.log_loss(y_true, y_pred, *, eps=1e-15, normalize=True, sample_weight=None, labels=None)