adding layer size calc to Neural Networkstutorial

It would have saved me a lot of wasted hours when moving from the tutorial to our own scenarios to explain the layer size calculation implicit in https://pytorch.org/tutorials/beginner/blitz/neural_networks_tutorial.html .

This was explained clearly in https://stackoverflow.com/questions/53784998/how-are-the-pytorch-dimensions-for-linear-layers-calculated

Could perhaps be a short addition to the "Note" section at the bottom?

(If that seems like a good solution / no one else gets to it, I should be able to get to this in mid-June)

Additionally: I don't think it belong in this tutorial tutorial, but there is a debugging tip that this common error message number is actually dependent on your batch size(!):
"RuntimeError: shape '[-1, 62500]' is invalid for input of size 11289600"
If you put in .view(-1, 11289600) the code will "run" further and then choke on the loss calculation; however, actually the original batch size issue has not been solved (as it should be set to 11289600 / batchsize), so attempting to address the loss error message does no good.

I'm not sure where this should go, but it's definitely pertinent for moving from the tutorial to real life (which the tutorials should give you the skills to do hopefully)

Start with the Neural Networks tutorial linked in the issue and review its Note section alongside the referenced Stack Overflow explanation. Clarify the layer-size calculation in the tutorial, then decide whether the batch-size debugging advice belongs there or should be documented elsewhere. Done means the tutorial explains the calculation accurately and the documentation scope is resolved.