Xlm R Finetuned Toxic Political Tweets Es

🚀 xlm-r-finetuned-toxic-political-tweets-es

This model, based on xlm-roberta-base, is fine-tuned on a dataset of tweets from the Spanish Congress of the Deputies. It predicts the political toxicity levels of these tweets.

🚀 Quick Start

This model is designed to detect the political toxicity of Spanish tweets from members of the Spanish Congress of the Deputies. It predicts two levels of toxicity: "Toxic" and "Very Toxic", each with a score between 0 and 1.

✨ Features

• Input: Trained on Spanish tweets authored by politicians in 2021. Performance may degrade on texts from other domains.
• Output: Predicts two signals of political toxicity: "Toxic" and "Very Toxic", with a score between 0 and 1 for each.

💻 Usage Examples

Basic Usage

>>> from transformers import pipeline
>>> text = "Es usted un auténtico impresentable, su señoría."
>>> pipe = pipeline("text-classification", model="Newtral/xlm-r-finetuned-toxic-political-tweets-es")
>>> pipe(text, return_all_scores=True)

[[{'label': 'toxic', 'score': 0.92560875415802},
  {'label': 'very toxic', 'score': 0.8310967683792114}]]

📚 Documentation

Intended uses & limitations

The model is intended for detecting the political toxicity of Spanish tweets from the Spanish Congress Deputies. For other uses, such as toxicity detection in film reviews, the results may not be reliable, and you may need to find another model for that specific purpose.

Training procedure

The pre-trained model was fine-tuned for sequence classification using the following hyperparameters, selected from a validation set:

The optimizer used was AdamW, and the loss optimized was binary cross-entropy with class weights proportional to the class imbalance.

📄 License

This project is licensed under the Apache-2.0 license.