Minor PR to Aerial.

Cope with the first chapter of Machine Learning with CoreML book. Later fixed link to cover book if repo with code from book.

Refactoring stupid snippet from the first chapter I jumped to Math.Floating where found about FloatingPoint protocol (Float, Double and Float80 conform to it). Float80 - that’s new for me! As well as BinaryFloatingPoint from there. Conclusion from browsing this file: instead of extending Float and/or Double with some functions one should consider defining this functions over FloatingPoint making new functionality available to all float types.

Just on time comes post about training YOLOv3 for custom objects. Hope to have time follow this tonight.