README.md

---
language: ko
license: mit
tags:
  - bart
  - grammar
---

# kogrammar-base

Dataset: 국립국어원 맞춤법 교정 말뭉치
<br>
<br>
Backbone Model: [kobart-base-v2](https://huggingface.co/gogamza/kobart-base-v2/blob/main/README.md)
<br>
GitHub Repo: [SOL1archive/KoGrammar](https://github.com/SOL1archive/KoGrammar)

## Train Method
전체 데이터셋 중 약 45%를 학습데이터로 활용하여 학습함.

## Metric
|BLEU-2|ROUGE-2 F1|
|-|-|
|77.8 %|55.0 %|