A Chinese word segment model based on BERT, F1-Score 97% - View it on GitHub