I don’t clearly understand why my answer is wrong !! Can someone please explain it to me ?
The softmax function re-scales all of the outputs so that their probabilities sum to 1.
That’s good if you’re only trying to identify one object, and you want the one with the highest probability.
But if you want to detect multiple objects, you want one output for each item, which ranges from 0 to 1, and then use a threshold to decide if it is true or false. That’s how sigmoid() works.