Why to Use BiLSTM when BERT itself is modeling the context?

macanv / BERT-BiLSTM-CRF-NER

Tensorflow solution of NER task Using BiLSTM-CRF model with Google BERT Fine-tuning And private Server services

https://github.com/macanv/BERT-BiLSMT-CRF-NER

4.68k stars 1.25k forks source link

Open mahdiabdollahpour opened 4 years ago

mahdiabdollahpour commented 4 years ago

Why Use BiLSTM when BERT itself is modeling the context? How much gain do we get going from BERT-CRF to BERT-BiLSTM-CRF