fancyzhx/ag_news
Viewer • Updated • 128k • 121k • 190
Modelo de clasificación de noticias en inglés basado en DistilBERT (distilbert-base-uncased),
entrenado mediante fine-tuning sobre el dataset AG News.
Clasifica noticias en 4 categorías temáticas:
| Categoría | Descripción |
|---|---|
| World News | International news and global politics |
| Sports | All sports coverage |
| Business | Economy, companies and markets |
| Science & Technology | Science, tech and innovation |
from transformers import pipeline
clasificador = pipeline('text-classification', model='raulbs7/ag-news-classifier')
result = classifier('Apple unveiled its latest iPhone at the annual keynote.')
print(result)
| Métrica | Valor |
|---|---|
| Accuracy | 0.8700 |
| F1 (weighted) | 0.8697 |
| Precision | 0.8716 |
| Recall | 0.8700 |
ag-news-classifier