---
language:
- "sr"
tags:
- "serbian"
- "token-classification"
- "pos"
- "dependency-parsing"
base_model: jerteh/gpt2-orao
datasets:
- "universal_dependencies"
license: "cc-by-sa-4.0"
pipeline_tag: "token-classification"
widget:
- text: "Да има сира и масла и моја би мати знала гибати гибаницу."
- text: "Da ima sira i masla i moja bi mati znala gibati gibanicu."
---

# gpt2-large-serbian-upos

## Model Description

This is a GPT-2 model in Serbian (Cyrillic and Latin) for POS-tagging and dependency-parsing, derived from [gpt2-orao](https://huggingface.co/jerteh/gpt2-orao). Every word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech) and [FEATS](https://universaldependencies.org/u/feat/).

## How to Use

```py
from transformers import pipeline
nlp=pipeline("upos","KoichiYasuoka/gpt2-large-serbian-upos",trust_remote_code=True,aggregation_strategy="simple")
```

or

```py
import esupar
nlp=esupar.load("KoichiYasuoka/gpt2-large-serbian-upos")
```

## See Also

[esupar](https://github.com/KoichiYasuoka/esupar): Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa/GPT models