Downloads · 30 days
76
31% of all-time downloads
deberta-v3-base-korean-ai-detect
hyun5ooo/deberta-v3-base-korean-ai-detect
deberta-v3-base-korean-ai-detect is a text classification model from hyun5ooo. Use it when you need a label for a piece of text. The card lists the license as apache-2.0.
All-time downloads
243
Public
Parameters
135M
541 MB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors541 MB · 100%
At a glance
- Task
- Text Classification
- License
- apache-2.0
- Model type
- deberta-v2
- Access
- Public
- Created
- Oct 8, 2025
- Updated
- Oct 14, 2025
- SHA
- f90625c8
Base models
Model card
From the Hugging Face model README
deberta-v3-base-korean-ai-detect
Model Details
2025 SW중심대학 디지털 경진대회 : AI부문 : 생성형 AI(LLM)와 인간 : 텍스트 판별 챌린지 수상작
이 모델은 한성대학교 SW중심대학사업단으로부터 후원 받은 Cloud GPU 로 학습되었습니다.
How to Get Started with the Model
from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained("team-lucid/deberta-v3-base-korean")
model = AutoModelForSequenceClassification.from_pretrained("hyun5ooo/deberta-v3-base-korean-ai-detect")
inputs = tokenizer("안녕, 세상!", return_tensors="pt")
outputs = model(**inputs)
Evaluation
2025 SW중심대학 디지털 경진대회 : AI부문 / Public 리더보드 기준
| Model | roc-auc |
|---|---|
| deberta base | 0.903 |
| deberta xlarge | 0.904 |
| mdeberta base | 0.894 |
| koelectra base | 0.894 |
| klue roberta large | 0.892 |
- Task
- Text Classification
- Type
- deberta-v2
- License
- apache-2.0
- Languages
- ko
- Created
- Oct 8, 2025
- Updated
- Oct 14, 2025