jisukim8873's picture
Update README.md
8b754e0
|
raw
history blame
685 Bytes
---
license: cc-by-4.0
language:
- en
- ko
tags:
- translation
---
## Model Details
* Model Description: Fine-tuned facebook/nllb-200-distilled-600M model
* Developed by: Jisu, Kim and Juhwan, Lee
* Model Type: Translation
* Language(s):
* Source Language: English
* Target Language: Korean
* License: CC-BY-4.0
## Dataset
* [AI-hub dataset](https://www.aihub.or.kr/)
## BLEU Score
* Before fine-tuning: 36.54
* After fine-tuning: 36.87
## Uses
This model can be used for translation and text-to-text generation
## Data Augmentation with backtranslation task
You can exercise korean data augmentation task with python package [KoTAN](https://github.com/KoJLabs/KoTAN/tree/main)