| 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) | |
| ``` | |