alpha-0.2.2

Vectorization Update!

Changed from using a neuron class and managing neurons in the layer to directly storing weights and inputs in the array. Achieved speeds almost 10x faster than old code and comparing to tensorflow is about a delta of 1 second for any datasize.

What's Changed

• Vectorize all multiplications and calculations by @neelthepatel8 in #17
• Save model by @cwcoogan in #18

Full Changelog: alpha0.2.1...alpha-0.2.2