---
library_name: grafon
language: sk
base_model: gerulata/slovakbert
pipeline_tag: text-to-speech
tags: [g2p, phonemes, ipa, grapheme-to-phoneme]
---

# sk

Contextual grapheme-to-phoneme for Slovak, trained with [grafon](https://github.com/thewh1teagle/grafon) on top of [gerulata/slovakbert](https://huggingface.co/gerulata/slovakbert).

```python
from grafon import Phonemizer

phonemize = Phonemizer.from_pretrained("grafon-g2p/sk")
print(phonemize('Ahoj svet, čo si myslíš o modeli?'))
```

Phonemes: `abcdefijklmnoprstuvxzŋɟɡɣɦɱɲʃʎʒʣʤʦʧˈː̯̩`. Anything outside the letters `aáäbcčdďeéfghiíjklĺľmnňoóôpqrŕsštťuúvwxyýzž` passes through as written.