BERTurk for Tobacco Control
This model is fine-tuned to identify tobacco-promoting tweets in Turkish.
It is based on a BERT model pretrained in Turkish: BERTurk (uncased).
Performance
| Metric | Value |
|---|---|
| Accuracy | 81.7% |
| PPV (Precision) | 81.1% |
| Sensitivity (Recall) | 87.8% |
| Specificity | 73.9% |
| F1 Score | 84.3% |
| Pearson’s phi (MCC) | 0.63 |
| Cohen’s kappa | 0.62 |
Usage
from simpletransformers.classification import ClassificationModel
model = ClassificationModel(
'bert',
'hkucukali/berturk-tobacco-control',
use_cuda=False,
num_labels=2,
)
prediction, probability = model.predict([tweet_text.lower()])
- Downloads last month
- 4
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.
Model tree for hkucukali/berturk-tobacco-control
Base model
dbmdz/bert-base-turkish-uncased