Let's start the exercises by calculating the loss function by hand. Don't do this exercise in PyTorch, it is important to first do it using only pen and paper (and a calculator).
We have the same example as before but now our object is actually a frog, and the predicted scores are
-1.2 for class
0.12 for class
1 (car) and
4.8 for class
What is the result of the softmax cross-entropy loss function?