shibing624
commited on
Commit
·
0c60c1e
1
Parent(s):
4a65218
Update README.md
Browse files
README.md
CHANGED
@@ -11,15 +11,17 @@ license: "apache-2.0"
|
|
11 |
# MacBERT for Chinese Spelling correction(macbert4csc) Model
|
12 |
中文拼写纠错模型
|
13 |
|
14 |
-
`macbert4csc-base-chinese` evaluate
|
15 |
|
16 |
-
|
|
|
|
|
|
|
17 |
|
18 |
-
模型在SIGHAN2015数据集达到SOTA。
|
19 |
|
20 |
## Usage
|
21 |
|
22 |
-
本项目开源在中文文本纠错项目:[pycorrector](https://github.com/shibing624/pycorrector),可支持
|
23 |
|
24 |
```python
|
25 |
from pycorrector.macbert.macbert_corrector import MacBertCorrector
|
|
|
11 |
# MacBERT for Chinese Spelling correction(macbert4csc) Model
|
12 |
中文拼写纠错模型
|
13 |
|
14 |
+
`macbert4csc-base-chinese` evaluate SIGHAN2015 test data:
|
15 |
|
16 |
+
- Char Level: precision=0.9372, recall=0.8640 F1=0.8991
|
17 |
+
- Sentence Level: precision:0.8264, recall:0.7366, f1:0.7789
|
18 |
+
|
19 |
+
由于训练使用的数据使用了SIGHAN2015的训练集(复现paper),在SIGHAN2015的测试集上达到SOTA水平。
|
20 |
|
|
|
21 |
|
22 |
## Usage
|
23 |
|
24 |
+
本项目开源在中文文本纠错项目:[pycorrector](https://github.com/shibing624/pycorrector),可支持macbert4csc模型,通过如下命令调用:
|
25 |
|
26 |
```python
|
27 |
from pycorrector.macbert.macbert_corrector import MacBertCorrector
|