Change the kernel size from 3x3 to the original 5x5

[matsui528]

This PR reverts the kernel size from 3x3 to the original 5x5 at the beginning of the 60 Minute Blitz.
https://pytorch.org/tutorials/beginner/blitz/neural_networks_tutorial.html#neural-networks

The kernel size was originally 5x5, but #515 changed it to 3x3 to avoid confusion on the FC parameter. It indeed makes sense.

However, it turned out that there are some drawbacks:

• The 3x3 kernel size contradicts an overview figure. https://pytorch.org/tutorials/_images/mnist.png
• The next tutorial that uses a similar architecture still has the original 5x5 kernel size. This inconsistency would confuse beginners. https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html#define-a-convolutional-neural-network

After some discussion (#503 (comment) and 6308024#commitcomment-34816516), it seems the best way is to revert the kernel size from 3x3 to the original 5x5.

[netlify]

Deploy preview for pytorch-tutorials-preview ready!

Built with commit 479ea7c

https://deploy-preview-1257--pytorch-tutorials-preview.netlify.app

[facebook-github-bot]

Hi @matsui528!

Thank you for your pull request and welcome to our community. We require contributors to sign our Contributor License Agreement, and we don't seem to have you on file.

In order for us to review and merge your code, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

If you have received this in error or have any questions, please contact us at cla@fb.com. Thanks!

[matsui528]

I found three duplicate PRs, meaning a strong demand for revising this inconsistent description. #1003 #859 #835

As this is the beginning of the tutorial, should the description be precise with a consistent relation to the overview figure?

[matsui528]

Related issues: #924 #839 #727 #724 #703 #694

[facebook-github-bot]

Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Facebook open source project. Thanks!