Tried to remove variants and increase face quality.
Second iteration of this model, hopefully this helps someone else create more consistent imagery as well.
