batch normalization layer
Created by: jngadiub
Start here with adding the class for computing batch normalization given some trained parameters as implemented here:
https://github.com/keras-team/keras/blob/master/keras/backend/tensorflow_backend.py#L1857
The gamma / sqrt(var + epsilon) factor is loaded directly as one value called "scale".
The hls-writer still needs to be updated to support this.