test
This commit is contained in:
xxupiano
parent
80f657949f
commit
67c872c161
62
README.md
62
README.md
|
|
@ -40,48 +40,60 @@ DeepKE 提供了多种知识抽取模型。
|
|||
## 在线演示
|
||||
演示的demo地址
|
||||
|
||||
1.NER
|
||||
1. RE
|
||||
|
||||
```
|
||||
REGULAR
|
||||
```
|
||||
```
|
||||
1.REGULAR
|
||||
|
||||
2.FEW-SHOT
|
||||
|
||||
3.DOCUMENT
|
||||
```
|
||||
|
||||
2.RE
|
||||
2. NER
|
||||
|
||||
1.REGULAR
|
||||
|
||||
2.FEW-SHOT
|
||||
|
||||
3.DOCUMENT
|
||||
```
|
||||
REGULAR
|
||||
```
|
||||
|
||||
3.AE
|
||||
3. AE
|
||||
|
||||
|
||||
## 快速上手
|
||||
|
||||
1.RE
|
||||
1. RE
|
||||
|
||||
数据为csv文件，样式范例为：
|
||||
数据为csv文件，样式范例为：
|
||||
|
||||
sentence|relation|head|head_offset|tail|tail_offset
|
||||
:---:|:---:|:---:|:---:|:---:|:---:
|
||||
《岳父也是爹》是王军执导的电视剧，由马恩然、范明主演。|导演|岳父也是爹|1|王军|8
|
||||
《九玄珠》是在纵横中文网连载的一部小说，作者是龙马。|连载网站|九玄珠|1|纵横中文网|7
|
||||
提起杭州的美景，西湖总是第一个映入脑海的词语。|所在城市|西湖|8|杭州|2
|
||||
| Sentence | Relation | Head | Head_offset | Tail | Tail_offset |
|
||||
| :----------------------------------------------------: | :------: | :--------: | :---------: | :--------: | :---------: |
|
||||
| 《岳父也是爹》是王军执导的电视剧，由马恩然、范明主演。 | 导演 | 岳父也是爹 | 1 | 王军 | 8 |
|
||||
| 《九玄珠》是在纵横中文网连载的一部小说，作者是龙马。 | 连载网站 | 九玄珠 | 1 | 纵横中文网 | 7 |
|
||||
| 提起杭州的美景，西湖总是第一个映入脑海的词语。 | 所在城市 | 西湖 | 8 | 杭州 | 2 |
|
||||
|
||||
具体流程请进入具体的README中，RE包括了以下三个子功能
|
||||
具体流程请进入详细的README中，RE包括了以下三个子功能
|
||||
|
||||
**[REGULAR](https://github.com/zjunlp/deepke/blob/test_new_deepke/example/re/regular/README.md)**
|
||||
**[REGULAR](https://github.com/zjunlp/deepke/blob/test_new_deepke/example/re/regular/README.md)**
|
||||
|
||||
FEW-SHOT
|
||||
FEW-SHOT
|
||||
|
||||
DOCUMENT
|
||||
DOCUMENT
|
||||
|
||||
2.NER
|
||||
2. NER
|
||||
|
||||
**[REGULAR](https://github.com/zjunlp/deepke/blob/test_new_deepke/example/ner/regular/README.md)**
|
||||
数据为txt文件，样式范例为：
|
||||
|
||||
3.AE
|
||||
| sentence | Person | Location | Organization | Miscellaneous |
|
||||
| :----------------------------------------------------------: | :----------------------------------: | :---------------: | :-------------------------: | :-------------------: |
|
||||
| Australian Tom Moody took six for 82 but Chris Adams, 123, and Tim O'Gorman, 109, took Derbyshire to 471 and a first innings lead of 233. | Tom Moody, Chris Adams, Tim O'Gorman | / | Derbysire | Australian |
|
||||
| Irene, a master student in Zhejiang University, Hangzhou, is traveling in Warsaw for Chopin Music Festival. | Irene | Hangzhou, Warsaw | Zhejiang University | Chopin Music Festival |
|
||||
| It was one o'clock when we left Lauriston Gardens and Sherlock Holmes led me to Metropolitan Police Service. | Sherlock Holmes | Lauriston Gardens | Metropolitan Police Service | / |
|
||||
|
||||
具体流程请进入详细的README中：
|
||||
|
||||
**[REGULAR](https://github.com/zjunlp/deepke/blob/test_new_deepke/example/ner/regular/README.md)**
|
||||
|
||||
3. AE
|
||||
|
||||
## 模型架构
|
||||
Deepke的架构图如下所示
|
||||
|
|
|
|||
|
|
@ -1,5 +1,17 @@
|
|||
# 快速上手
|
||||
|
||||
## 环境依赖
|
||||
|
||||
> python >= 3.7
|
||||
|
||||
- pytorch-transformers==1.2.0
|
||||
- torch==1.2.0
|
||||
- seqeval==0.0.5
|
||||
- tqdm==4.31.1
|
||||
- nltk==3.4.5
|
||||
|
||||
|
||||
|
||||
## 克隆代码
|
||||
|
||||
```
|
||||
|
|
@ -8,34 +20,34 @@ git clone git@github.com:zjunlp/DeepKE.git
|
|||
|
||||
|
||||
|
||||
## 配置环境
|
||||
## 使用pip安装
|
||||
|
||||
创建python虚拟环境（python>=3.7）
|
||||
首先创建python虚拟环境，再进入虚拟环境
|
||||
|
||||
安装依赖库
|
||||
|
||||
```
|
||||
pip install -r requirements.txt
|
||||
```
|
||||
- 安装依赖：`pip install -r requirements.txt`
|
||||
|
||||
|
||||
|
||||
## 使用工具
|
||||
## 使用数据进行训练预测
|
||||
|
||||
先进行训练，训练后的模型参数保存在out_ner文件夹中
|
||||
- 存放数据：在`data`文件夹下存放数据集。主要有三个文件：
|
||||
- `train.txt`：存放训练数据集
|
||||
- `valid.txt`：存放验证数据集
|
||||
- `test.txt`：存放测试数据集
|
||||
|
||||
```
|
||||
python run.py --data_dir=data/ --bert_model=bert-base-cased --task_name=ner --output_dir=out_ner --max_seq_length=128 --do_train --num_train_epochs 5 --do_eval --warmup_proportion=0.1
|
||||
```
|
||||
- 先进行训练，训练后的模型参数保存在out_ner文件夹中
|
||||
|
||||
再进行预测<br>
|
||||
```
|
||||
python run.py --data_dir=data/ --bert_model=bert-base-cased --task_name=ner --output_dir=out_ner --max_seq_length=128 --do_train --num_train_epochs 5 --do_eval --warmup_proportion=0.1
|
||||
```
|
||||
|
||||
执行以下命令运行示例
|
||||
- 再进行预测
|
||||
|
||||
- 执行以下命令运行示例`python predict.py`
|
||||
|
||||
- 如果需要指定NER的文本，可以利用--text参数指定，如：
|
||||
|
||||
````
|
||||
python predict.py --text="It was one o'clock when we left Lauriston Gardens and Sherlock Holmes led me to Metropolitan Police Service.."
|
||||
````
|
||||
|
||||
```
|
||||
python predict.py
|
||||
```
|
||||
如果需要指定NER的文本，可以利用--text参数指定
|
||||
```
|
||||
python predict.py --text="Irene, a master student in Zhejiang University, Hangzhou, is traveling in Warsaw for Chopin Music Festival."
|
||||
```
|
||||
|
|
|
|||
|
|
@ -19,7 +19,7 @@ git clone git@github.com:zjunlp/DeepKE.git
|
|||
```
|
||||
### 使用pip安装
|
||||
|
||||
首先创建python虚拟环境，再进入虚拟环境,然
|
||||
首先创建python虚拟环境，再进入虚拟环境
|
||||
|
||||
- 安装依赖: ```pip install -r requirements.txt```
|
||||
|
||||
|
|
|
|||
Loading…
Reference in New Issue