Hello. While predicting probability using sigmoid activation output will be in the form of
Prediction. Label
Tensor([0.95,0.123]) tensor([1.,0.])
Tensor([0.96,0.42]) tensor([0.,1.])
…so on
Can u please tell me why that 2 probability values are displayed?. And in Label why once it well display 1,0 fit some and 0,1 for some images?
Because you have 2 classes.
Yes. Now how can I classify using these values?
Tell me how can I classify images that belongs to class 0 and also class 1?
Typically you would classify with whatever probability is highest.
Tensor([0.95,0.123]) tensor([1.,0.])
Tensor([0.96,0.42]) tensor([0.,1.])
Would mean both examples are classified as class 0. Second one has ground truth label of class 1.
I didn’t understood Sir. How it is classified as class 0. Because in second case probability value is more for label 0. In first case it is opposite .