Downloads · 30 days
8
17% of all-time downloads
JOhyeongi/vet-kmbert-cross-encoder
vet-kmbert-cross-encoder is a text classification model from JOhyeongi. Use it when you need a label for a piece of text. It is set up for sentence-transformers. The card lists the license as mit.
수의학 도메인에 특화된 한국어 Cross-Encoder 모델입니다. RAG 시스템의 Reranking 단계에서 사용됩니다.
Downloads · 30 days
8
17% of all-time downloads
All-time downloads
48
Public
Parameters
98.7M
395 MB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors395 MB · 100%
From the Hugging Face model README
수의학 도메인에 특화된 한국어 Cross-Encoder 모델입니다. RAG 시스템의 Reranking 단계에서 사용됩니다.
| Metric | Score |
|---|---|
| Accuracy | ~68% |
| F1-Score | ~0.72 |
| Precision | ~0.71 |
| Recall | ~0.73 |
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch
# 모델 로드
model_name = "JOhyeongi/vet-kmbert-cross-encoder"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)
# 추론
query = "강아지가 구토를 해요."
document = "강아지 구토의 원인은 다양합니다..."
inputs = tokenizer(
[[query, document]],
padding=True,
truncation=True,
return_tensors="pt",
max_length=512
)
with torch.no_grad():
logits = model(**inputs).logits
probs = torch.softmax(logits, dim=1)
score = probs[0][1].item() # Relevance score
print(f"Relevance Score: {score:.4f}")
이 모델은 다음 프로젝트의 일부입니다:
Epochs: 3
Batch Size: 8
Learning Rate: 2e-5
Max Length: 512
Optimizer: AdamW
Weight Decay: 0.01
Warmup Steps: 500
MIT License
@misc{vet-kmbert-cross-encoder,
title={Vet KM-BERT Cross-Encoder: Korean Veterinary RAG System},
author={Catholic University},
year={2025},
publisher={Hugging Face},
url={https://huggingface.co/JOhyeongi/vet-kmbert-cross-encoder}
}