I was looking at https://pytorch.org/tutorials/beginner/blitz/neural_networks_tutorial.html#sphx-glr-beginner-blitz-neural-networks-tutorial-py. For the network it defines, the FC3_weight connects F6_layer and output_layer without ReLU, and then put into crossentropy loss.
But according to the illustration of the network, also the LeCun paper (http://vision.stanford.edu/cs598_spring07/papers/Lecun98.pdf), the output layer has been Euclidean Radial Basis Function unit, rather than further pushes into the crossentropy loss. So this illustration here does not match with the network defined below right?
It is kind of confusing. Thank you very much.