Hello, I’m trying to use the inceptionv3 pytorch model but the outputs just seem odd. While in a ResNet my outputs look like this:
tensor([[0.3754],
[0.5348],
[0.4032],
[0.3661],
[0.4373],
[0.3900],
[0.4651],
[0.4704],
[0.4020],
[0.3750],
[0.5548],
[0.4493],
[0.3471],
...
on Inception I get this:
tensor([[-2.0164],
[-2.2767],
[-3.5188],
[-2.0595],
[ 0.1911],
[-2.0991],
[-0.4889],
[-1.2856],
[-1.0438],
[-3.2926],
[-2.7320],
[-0.4018],
[-0.8456],
[ 1.7160],
[ 2.0525],
[-4.7007],
[-2.5639],
[-0.3776],
[-0.1715],
[-0.7751],
[-2.6474],
[-0.4403],
[-1.1878],
[-0.8396],
[-1.7131],
[-0.5675],
[-0.7635],
I can’t understand the values amplitude and how to I get something binary from here. Is there any post processing required after Inception or it is just my implementation that is incorrect?
Here’s my implementation of the InceptionV3 model.
Regards