Hate-speech-CNERG
/

indic-abusive-allInOne-MuRIL

Text Classification

Inference Endpoints

Model card Files Files and versions Community

Edit model card

YAML Metadata Error: "language[2]" with value "hi-en" is not valid. It must be an ISO 639-1, 639-2 or 639-3 code (two/three letters), or a special value like "code", "multilingual". If you want to use BCP-47 identifiers, you can specify them in language_bcp47.

YAML Metadata Error: "language[3]" with value "ka-en" is not valid. It must be an ISO 639-1, 639-2 or 639-3 code (two/three letters), or a special value like "code", "multilingual". If you want to use BCP-47 identifiers, you can specify them in language_bcp47.

YAML Metadata Error: "language[4]" with value "ma-en" is not valid. It must be an ISO 639-1, 639-2 or 639-3 code (two/three letters), or a special value like "code", "multilingual". If you want to use BCP-47 identifiers, you can specify them in language_bcp47.

YAML Metadata Error: "language[6]" with value "ta-en" is not valid. It must be an ISO 639-1, 639-2 or 639-3 code (two/three letters), or a special value like "code", "multilingual". If you want to use BCP-47 identifiers, you can specify them in language_bcp47.

YAML Metadata Error: "language[8]" with value "ur-en" is not valid. It must be an ISO 639-1, 639-2 or 639-3 code (two/three letters), or a special value like "code", "multilingual". If you want to use BCP-47 identifiers, you can specify them in language_bcp47.

This model is used detecting abusive speech in Bengali, Devanagari Hindi, Code-mixed Hindi, Code-mixed Kannada, Code-mixed Malayalam, Marathi, Code-mixed Tamil, Urdu, Code-mixed Urdu, and English languages. The allInOne in the name refers to the Joint training/Cross-lingual training, where the model is trained using all the languages data. It is finetuned on MuRIL model. The model is trained with learning rates of 2e-5. Training code can be found at this url

LABEL_0 :-> Normal

LABEL_1 :-> Abusive

For more details about our paper

Mithun Das, Somnath Banerjee and Animesh Mukherjee. "Data Bootstrapping Approaches to Improve Low Resource Abusive Language Detection for Indic Languages". Accepted at ACM HT 2022.

Please cite our paper in any published work that uses any of these resources.

@article{das2022data,
  title={Data Bootstrapping Approaches to Improve Low Resource Abusive Language Detection for Indic Languages},
  author={Das, Mithun and Banerjee, Somnath and Mukherjee, Animesh},
  journal={arXiv preprint arXiv:2204.12543},
  year={2022}
}

Downloads last month: 130

Safetensors

Model size

238M params

Tensor type

I64

·

F32

·

Inference Examples

Text Classification

This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.