README.md

---
language:
- zh
tags:
- named entity recognition
license: apache-2.0
datasets:
- CMeEE
metrics:
- micro f1
---

# Global Pointer实体抽取模型

## 模型介绍

+ 数据集：CMeEE中文医学文本命名实时识别

+ 模型方法：[Efficient GlobalPointer：少点参数，多点效果](https://spaces.ac.cn/archives/8877)

## 使用方法

```commandline
pip install lightningnlp
```

```python
from pprint import pprint
from lightningnlp.task.named_entity_recognition import NerPipeline

pipline = NerPipeline(model_name_or_path="xusenlin/cmeee-global-pointer", model_name="global-pointer", model_type="bert")
text = "可伴发血肿或脑梗死而出现局灶性神经体征，如肢体瘫痪及颅神经异常等。"
pprint(pipline(text))
```