sagorsarker's picture
Add multilingual to the language tag (#1)
9ba9cb1
---
language:
- hi
- en
- multilingual
license: mit
tags:
- codeswitching
- hindi-english
- language-identification
datasets:
- lince
---
# codeswitch-hineng-lid-lince
This is a pretrained model for **language identification** of `hindi-english` code-mixed data used from [LinCE](https://ritual.uh.edu/lince/home)
This model is trained for this below repository.
[https://github.com/sagorbrur/codeswitch](https://github.com/sagorbrur/codeswitch)
To install codeswitch:
```
pip install codeswitch
```
## Identify Language
* **Method-1**
```py
from transformers import AutoTokenizer, AutoModelForTokenClassification, pipeline
tokenizer = AutoTokenizer.from_pretrained("sagorsarker/codeswitch-hineng-lid-lince")
model = AutoModelForTokenClassification.from_pretrained("sagorsarker/codeswitch-hineng-lid-lince")
lid_model = pipeline('ner', model=model, tokenizer=tokenizer)
lid_model("put any hindi english code-mixed sentence")
```
* **Method-2**
```py
from codeswitch.codeswitch import LanguageIdentification
lid = LanguageIdentification('hin-eng')
text = "" # your code-mixed sentence
result = lid.identify(text)
print(result)
```