setfit_camembert_20250808_140130_rtx4070_optimized

This is a SetFit model fine-tuned for French news classification as part of the AFK.live project.

Model Details

• Base Model: CamemBERT (camembert-base)
• Task: Multi-class text classification for news articles
• Language: French
• Framework: SetFit

Training Details

• Training Date: 2025-08-08
• Accuracy: 0.8700

Categories

The model classifies news articles into the following categories:

1. Politique nationale française ou belge
2. Politique internationale
3. Économie, Fiscalité et Mouvements sociaux
4. Technologie
5. Société, Média, Education, Histoire, Justice -
6. Science et Santé
7. Culture, Cuisine, Voyages
8. Sport
9. Environnement
10. Faits divers, nécrologie
11. Outlier
12. Actualité locale
13. Pub

Usage

from setfit import SetFitModel

# Load the model
model = SetFitModel.from_pretrained("DyePop/{model_name}")

# Make predictions
predictions = model.predict([
    "Le président a annoncé de nouvelles mesures économiques",
    "L'équipe de France remporte la victoire"
])

Limitations

• Trained specifically on French news articles (France, Walloon Belgium)
• Performance may vary on informal or social media text
• Categories are based on traditional news taxonomy

Citation

If you use this model, please cite the AFK.live project.