Tagger

Tagger could use as named entity recognized(NER), part of speech tagging(POS), word segmentation

This repo use torch and RNN-CRF to do tag job

这个项目是Infinity-Future Chatbot、Dialog System、NLP 等项目的一个子项目，目的在于构建通用、易用的中英文Tagger工具，主要工作可能是命名实体识别 (NER)

Document

Document Online

Score

tests/ner_glove_charcnn.py could reach CONLL2003 testb F1 >= 90.05, using GloVe.6B 100 dimension pretrained-vector

tests/segmentation_glove.py could reach MSR dataset test F1 >= 95.3, using Chinese character GloVe, pretrained-vector by author of this repo

Generate

Install documentation dependencies

pip3 install sphinx sphinx_rtd_theme --upgrade --user

run the script to generate docs

scripts/gen_doc.sh

Test

python3 -m tests