File size: 914 Bytes
ef606a8 0915d84 ef606a8 5175d95 ef606a8 5175d95 ef606a8 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 |
---
language:
- "ko"
tags:
- "korean"
- "token-classification"
- "pos"
- "dependency-parsing"
base_model: KoichiYasuoka/roberta-base-korean-hanja
datasets:
- "universal_dependencies"
license: "cc-by-sa-4.0"
pipeline_tag: "token-classification"
widget:
- text: "홍시 맛이 나서 홍시라 생각한다."
- text: "紅柹 맛이 나서 紅柹라 生覺한다."
---
# roberta-base-korean-ud-goeswith
## Model Description
This is a RoBERTa model pre-trained on Korean texts for POS-tagging and dependency-parsing (using `goeswith` for subwords), derived from [roberta-base-korean-hanja](https://huggingface.co/KoichiYasuoka/roberta-base-korean-hanja).
## How to Use
```py
from transformers import pipeline
nlp=pipeline("universal-dependencies","KoichiYasuoka/roberta-base-korean-ud-goeswith",trust_remote_code=True,aggregation_strategy="simple")
print(nlp("홍시 맛이 나서 홍시라 생각한다."))
```
|