Use BERT to get sentence and tokens embedding in an easier way (github.com)

BERT was one of the most exciting NLP papers published in 2018. Instead of using BERT to build an end-to-end model, using word representations from BERT can help you improve your model performance a lot, but save a lot of computing resources.

My idea has been implemented as a bert-embedding and published on Github. check: https://github.com/imgarylai/bert-embedding