Correct classification results but strange probabilities output

C++
1

I trained a classification model using Pytorch .

The inference result is correct but the probabilities output is strange , for example I get :

Prob = [4.5566592, -0.49383235, -4.1694036] . the probabilities are not in the interval [0,1] .

Is this correct ? or I am missing something

Thank you.

2

how would we know ?
we can neither see your pytorch model, nor your opencv code.
you also need to check, if your pre/post pocessing is the same.
it also depends on your last activation function (which is ?)