Multitask Classifier return probability assay

#1

Multitask Classifier return probability array

array([[[0.50144464, 0.49855536]],
[[0.50288945, 0.4971106 ]]], dtype=float32)

which one is for class 1 and which one is for class 0?

Moreover, in multitask classifier how to compute specificity?

Thank you

#2

In the example, 0.501… would be class 0 and 0.498 would be class 1. You can compute specificity using dc.metrics I believe (or if not using a metric pulled from sklearn)