an upload of my exploration of the effects of 8 bit quantization on how it effects error in weights and ranges in deep neural networks