Hardware Acceleration of Fully Quantized BERT for Efficient Natural Language Processing | IEEE Conference Publication | IEEE Xplore