---
license: openrail++
language:
- uk
widget:
- text: Ти неймовірна!
datasets:
- ukr-detect/ukr-toxicity-dataset
---

## Binary toxicity classifier for Ukrainian.

This is the fine-tuned on the downstream task ["distilbert-base-multilingual-cased"](https://huggingface.co/distilbert-base-multilingual-cased) instance.

The evaluation metrics for binary toxicity classification are: 

**Precision**: 0.9310
**Recall**: 0.9300
**F1**: 0.9300

The training and evaluation data will be clarified later.

## How to use
```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification

# load tokenizer and model weights
tokenizer = AutoTokenizer.from_pretrained('dardem/mdistilbert-base-cased-uk-toxicity')
model = AutoModelForSequenceClassification.from_pretrained('dardem/mdistilbert-base-cased-uk-toxicity')

# prepare the input
batch = tokenizer.encode('Ти неймовірна!', return_tensors='pt')

# inference
model(batch)
```

## Citation

```
@article{dementieva2024toxicity,
  title={Toxicity Classification in Ukrainian},
  author={Dementieva, Daryna and Khylenko, Valeriia and Babakov, Nikolay and Groh, Georg},
  journal={arXiv preprint arXiv:2404.17841},
  year={2024}
}
```