🌞 2-训练NER模型
1. dataset source: [chineseBLEU](https://github.com/alibaba-research/ChineseBLUE)
2. we first preprocess the dataset so that each sample contains a sequence of tokens and a sequence of tags [[colab](https://colab.research.google.com/gist/leoxiang66/a50ea0713cd9e99daa4734f1e61ce5c8/build-ner-dataset.ipynb)]
3. then we fine-tune the chinese BERT on this task for the NER task [[colab](https://colab.research.google.com/gist/leoxiang66/6510451934d15703a80c73b401c87a1a/finetune-chinese-bert-ner-biomedical.ipynb)]
4. to use the model, see [[inference example](https://colab.research.google.com/gist/leoxiang66/f77e7b6d893276a6bb68ffd7951f2ffa/untitled46.ipynb)]