yumath/bertNER
A Chinese named entity recognition system that fine-tunes BERT as an embedding layer with a BiLSTM+CRF decoder.
Not currently ranked — collecting fresh signals.
star history
This project applies Google BERT as a feature extraction layer before a BiLSTM+CRF architecture for Chinese named entity recognition. The BERT embeddings are fine-tuned during training on NER data from ChineseNER. It achieves approximately 93.68% F1 on the test set, improving over the non-BERT baseline by over two percentage points.
Frequently asked
- What is yumath/bertNER?
- A Chinese named entity recognition system that fine-tunes BERT as an embedding layer with a BiLSTM+CRF decoder.
- Is bertNER open source?
- Yes — yumath/bertNER is an open-source project tracked on heatdrop.
- What language is bertNER written in?
- yumath/bertNER is primarily written in Python.
- How popular is bertNER?
- yumath/bertNER has 458 stars on GitHub.
- Where can I find bertNER?
- yumath/bertNER is on GitHub at https://github.com/yumath/bertNER.