KoGrammar
Collection
3 items
•
Updated
kogrammar-distil
Dataset: 국립국어원 맞춤법 교정 말뭉치
Backbone Model:
Contributors:
- ChangGeon Kang: NeverAsking
- Subin Park: SOL1archive
GitHub Repo:
Train Method
전체 데이터셋 중 약 67.5%를 학습데이터로 활용하여 학습함.
SFT Distillation을 이용해 kogrammar-base 모델의 Decoder Layer를 6개에서 3개로 줄여 다시 학습시킴.
Metric
| BLEU-2 | ROUGE-2 F1 |
|---|---|
| 77.8 % | 55.0 % |
- Downloads last month
- 18
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.