FISA-conclave / klue-roberta-news-sentiment

huggingface.co
Total runs: 16.9K
24-hour runs: 0
7-day runs: -1.3K
30-day runs: -1.1K
Model's Last Updated: April 06 2026
text-classification

Introduction of klue-roberta-news-sentiment

Model Details of klue-roberta-news-sentiment

KLUE-RoBERTa 뉴스 기사 기업 감정분석 모델

모델 설명

이 모델은 뉴스 기사 속 특정 기업에 대한 감정(호재/악재/중립)을 분석 하기 위해 fine-tuning된 KLUE-RoBERTa 모델입니다.

사용 목적
  • 뉴스 기사에서 특정 기업에 대한 긍정/부정/중립 감정 자동 분류
  • 금융 뉴스 감정 분석
  • 기업 평판 모니터링
레이블
  • 0 : negative (악재/부정) - 해당 기업에 대한 부정적 내용
  • 1 : neutral (중립) - 해당 기업에 대한 중립적 내용
  • 2 : positive (호재/긍정) - 해당 기업에 대한 긍정적 내용
성능
Metric Score
Accuracy 0.8426
F1-Macro 0.8468
F1-Weighted 0.8422
하이퍼파라미터 (Optuna로 최적화)
{
  "learning_rate": 9.78310992630157e-06,
  "num_train_epochs": 8,
  "weight_decay": 0.06436845335086991,
  "warmup_ratio": 0.10859899755289561,
  "per_device_train_batch_size": 32
}
사용 방법
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

# 모델 로드
model_name = "FISA-conclave/klue-roberta-news-sentiment"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

# 예측
text = "삼성전자의 3분기 실적이 시장 예상을 크게 상회했다."
inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=128)

with torch.no_grad():
    outputs = model(**inputs)
    probs = torch.softmax(outputs.logits, dim=1)[0]
    pred = torch.argmax(probs).item()

labels = {0: "negative", 1: "neutral", 2: "positive"}
print(f"예측: {labels[pred]} ({probs[pred]:.2%})")
학습 데이터
  • 총 샘플 수 : 9,992개
  • 출처 :
    • finance_sentiment_corpus
    • korfin-asc
    • twice_kr_fin
베이스 모델
인용
@misc{klue-roberta-news-sentiment,
  author = {Tobykim},
  title = {KLUE-RoBERTa News Sentiment Analysis},
  year = {2024},
  publisher = {HuggingFace},
  howpublished = {\url{https://huggingface.co/FISA-conclave/klue-roberta-news-sentiment}}
}
라이센스

Apache 2.0

Runs of FISA-conclave klue-roberta-news-sentiment on huggingface.co

16.9K
Total runs
0
24-hour runs
351
3-day runs
-1.3K
7-day runs
-1.1K
30-day runs

More Information About klue-roberta-news-sentiment huggingface.co Model

More klue-roberta-news-sentiment license Visit here:

https://choosealicense.com/licenses/apache-2.0

klue-roberta-news-sentiment huggingface.co

klue-roberta-news-sentiment huggingface.co is an AI model on huggingface.co that provides klue-roberta-news-sentiment's model effect (), which can be used instantly with this FISA-conclave klue-roberta-news-sentiment model. huggingface.co supports a free trial of the klue-roberta-news-sentiment model, and also provides paid use of the klue-roberta-news-sentiment. Support call klue-roberta-news-sentiment model through api, including Node.js, Python, http.

klue-roberta-news-sentiment huggingface.co Url

https://huggingface.co/FISA-conclave/klue-roberta-news-sentiment

FISA-conclave klue-roberta-news-sentiment online free

klue-roberta-news-sentiment huggingface.co is an online trial and call api platform, which integrates klue-roberta-news-sentiment's modeling effects, including api services, and provides a free online trial of klue-roberta-news-sentiment, you can try klue-roberta-news-sentiment online for free by clicking the link below.

FISA-conclave klue-roberta-news-sentiment online free url in huggingface.co:

https://huggingface.co/FISA-conclave/klue-roberta-news-sentiment

klue-roberta-news-sentiment install

klue-roberta-news-sentiment is an open source model from GitHub that offers a free installation service, and any user can find klue-roberta-news-sentiment on GitHub to install. At the same time, huggingface.co provides the effect of klue-roberta-news-sentiment install, users can directly use klue-roberta-news-sentiment installed effect in huggingface.co for debugging and trial. It also supports api for free installation.

klue-roberta-news-sentiment install url in huggingface.co:

https://huggingface.co/FISA-conclave/klue-roberta-news-sentiment

Url of klue-roberta-news-sentiment

klue-roberta-news-sentiment huggingface.co Url

Provider of klue-roberta-news-sentiment huggingface.co

FISA-conclave
ORGANIZATIONS