Roberta Base Ca Finetuned Cyberbullying Catalan

🚀 roberta-base-ca-finetuned-cyberbullying-catalan

This model is a fine - tuned version of [BSC - TeMU/roberta - base - ca](https://huggingface.co/BSC - TeMU/roberta - base - ca) to detect cyberbullying in Catalan, trained on data scraped from various social networks.

🚀 Quick Start

This model is a fine - tuned version of [BSC - TeMU/roberta - base - ca](https://huggingface.co/BSC - TeMU/roberta - base - ca) on the dataset generated by scraping all social networks (Twitter, Youtube ...) to detect cyberbullying in Catalan.

It achieves the following results on the evaluation set:

• Loss: 0.1508
• Accuracy: 0.9665

✨ Features

• Fine - tuned from [BSC - TeMU/roberta - base - ca](https://huggingface.co/BSC - TeMU/roberta - base - ca) for Catalan cyberbullying detection.
• High accuracy (0.9665) on the evaluation set.

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

Fast usage with pipelines:

from transformers import pipeline

model_path = "JonatanGk/roberta-base-ca-finetuned-ciberbullying-catalan"
bullying_analysis = pipeline("text-classification", model=model_path, tokenizer=model_path)

bullying_analysis(
    "Des que et vaig veure m'en vaig enamorar de tu."
    )
    
# Output:
[{'label': 'Not_bullying', 'score': 0.9996786117553711}]

bullying_analysis(
    "Ets tan lletja que et donaven de menjar per sota la porta."
    )
    
# Output:
[{'label': 'Bullying', 'score': 0.9927878975868225}]
    

You can try this model in Google Colab:

📚 Documentation

Training and evaluation data

I used the concatenation of multiple datasets generated by scraping social networks (Twitter, Youtube, Discord ...) to fine - tune this model. The total number of sentence pairs is above 410k sentences. A similar training method was used at [roberta - base - bne - finetuned - cyberbullying - spanish](https://huggingface.co/JonatanGk/roberta - base - bne - finetuned - cyberbullying - spanish)

Training procedure

Framework versions

• Transformers 4.10.3
• Pytorch 1.9.0+cu102
• Datasets 1.12.1
• Tokenizers 0.10.3

📄 License

No license information is provided in the original document.

📚 Citation

@inproceedings{armengol-estape-etal-2021-multilingual,
    title = "Are Multilingual Models the Best Choice for Moderately Under-resourced Languages? {A} Comprehensive Assessment for {C}atalan",
    author = "Armengol-Estap{\'e}, Jordi  and
      Carrino, Casimiro Pio  and
      Rodriguez-Penagos, Carlos  and
      de Gibert Bonet, Ona  and
      Armentano-Oller, Carme  and
      Gonzalez-Agirre, Aitor  and
      Melero, Maite  and
      Villegas, Marta",
    booktitle = "Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021",
    month = aug,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.findings-acl.437",
    doi = "10.18653/v1/2021.findings-acl.437",
    pages = "4933--4946",
}

Special thanks to Manuel Romero/@mrm8488 as my mentor & R.C.

Created by Jonatan Luna | LinkedIn