Add support for convolutional layers

[milancurcic]

This issue tracks the progress on the support for convolutional layers. Ideally will be using GH Projects for this but that's on an Org level; I don't want to pollute the modern-fortran Org with it, so when we transfer to github.com/neural-fortran Org, will set up some projects.

• Refactor for convnets #58
• conv2d forward pass #60
• Implement a max-pooling layer #63
• Implement a flatten layer #68
• conv2d backward pass #61
• Implement maxpool2d backward pass #69
• Add a MNIST training example using conv2d layers #62