Skip to content

Latest commit

 

History

History
33 lines (26 loc) · 622 Bytes

README.md

File metadata and controls

33 lines (26 loc) · 622 Bytes
Raw

Chinese NER

Model

Bert+FFN

Requirements

The main requirements are:

  • python 3.6
  • torch 1.7.0
  • tqdm
  • transformers 3.5.1

Dataset

Download CLUNER and put it under ./datasets/CLUENER.

Usage

  • Get pre-trained BERT model Download bert-base-chinese and put it under ./pretrained.

  • Train

python run.py --train train
  • Evaluate on the Test
python run.py --train predict

Result(F1)

dev test
76.2 77.5