Are you sure you want to delete this task? Once this task is deleted, it cannot be recovered.
You can not select more than 25 topics
Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
xiongying
1d1f15f582
|
1 year ago | |
|---|---|---|
| data | 2 years ago | |
| layer | 2 years ago | |
| model | 2 years ago | |
| process_code | 2 years ago | |
| utils | 2 years ago | |
| .gitignore | 2 years ago | |
| 1.png | 2 years ago | |
| 1.txt | 2 years ago | |
| README.en.md | 2 years ago | |
| README.md | 1 year ago | |
| infer.sh | 2 years ago | |
| main.py | 2 years ago | |
| meizhou1.txt | 2 years ago | |
| meizhou2.txt | 2 years ago | |
| meizhou3.txt | 2 years ago | |
| meizhou4.txt | 2 years ago | |
| meizhou5.txt | 2 years ago | |
| meizhou_all.txt | 2 years ago | |
| meizhou_dev.json | 2 years ago | |
| run.sh | 2 years ago | |
| run_ch.sh | 2 years ago | |
| run_en.sh | 2 years ago | |
| test.json | 2 years ago | |
| test.py | 2 years ago | |
| test.txt | 2 years ago | |
| train.txt | 2 years ago | |
README.md
KG-NER
模型介绍
knowledge-enhanced flat NER(named entity recognition)。对于中文则是引入知识图谱概念以及词典信息，对于英文可以引入知识图谱概念和短语信息。主要通过构建图来进行传递信息。
环境
cudnn 9.2
python 3.6
pytorch 1.6.0
transformers 3.4.0
pytorch-pretrained-bert 0.6.2
参数文件及设置
# 文件
utils/config.py
# 具体一些设置
数据相关
词典或者概念形式：
当为中文时，设置语言为“CN”，词语中的字和字之间没有操作，如“中国”
当为英文时或者西文时，设置语言为“EN”，概念或者短语的token之间用“&”分割，例如“bacteroides fragilis”在词典里为“bacteroides&fragilis”。
训练数据格式：
字符 label 起始位置 终止位置
主 O 0 1
因 O 2 4
腰 B-身体部位 5 9
部 I-身体部位 10 13
疼 B-症状和体征 14 17
痛 I-症状和体征 18 20
1 O 21 23
小 O 24 26
时 O 27 29
于 O 30 32
2016 O 33 35
每行用“\t”分隔，句子之间按照空行分隔开。
训练
run.sh
推理
infer.sh
reference
Xiong Y, Peng H, Xiang Y, et al. Leveraging Multi-source knowledge for Chinese clinical named entity recognition via relational graph convolutional network[J]. Journal of Biomedical Informatics, 2022, 128: 104035.
