merrybabyxmas/mindcast-emotion-sc-only

Mindcast Emotion Classifier

Model Description

한국어 텍스트에서 6가지 감정(분노, 슬픔, 불안, 상처, 당황, 기쁨)을 분류하는 모델입니다. 일반 댓글과 비꼬는 댓글(sarcasm) 모두를 학습하여 다양한 문맥에서 감정을 정확하게 인식합니다.

이 모델은 LoRA (Low-Rank Adaptation)를 사용하여 효율적으로 파인튜닝되었으며, 최종적으로 base model과 merge되어 배포되었습니다.

Training Date: 2025-12-15

Performance

Test Set Results

Metric	Score
Accuracy	0.2111
F1 Score (Macro)	0.1663
F1 Score (Weighted)	0.2473

Confusion Matrix

[[56 68 31 66 36 58]
 [ 6 10  6 10  6  4]
 [ 2  8  2  5  4  0]
 [ 1  0  0  1  1  0]
 [13 25  5 13 16 29]
 [ 3 10  4  7  5 29]]

Detailed Classification Report

              precision    recall  f1-score   support

          분노     0.6914    0.1778    0.2828       315
          슬픔     0.0826    0.2381    0.1227        42
          불안     0.0417    0.0952    0.0580        21
          상처     0.0098    0.3333    0.0190         3
          당황     0.2353    0.1584    0.1893       101
          기쁨     0.2417    0.5000    0.3258        58

   micro avg     0.2111    0.2111    0.2111       540
   macro avg     0.2171    0.2505    0.1663       540
weighted avg     0.4814    0.2111    0.2473       540

Training Details

Hyperparameters

Hyperparameter	Value
Base Model	klue/roberta-base
Batch Size	64
Epochs	50
Learning Rate	0.0001
Warmup Ratio	0.1
Weight Decay	0.01
LoRA r	8
LoRA alpha	16
LoRA dropout	0.05

Training Data

• Train samples: 1420 (from 2 files)
• Valid samples: 158
• Test samples: 540 (from 2 files)
• Number of labels: 6
• Labels: 분노, 슬픔, 불안, 상처, 당황, 기쁨

Usage

Installation

pip install transformers torch

Quick Start

from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline

# Load model
model_name = "merrybabyxmas/mindcast-emotion-sc-only"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

# Create pipeline
classifier = pipeline("text-classification", model=model, tokenizer=tokenizer)

# Predict
text = "오늘 정말 기분이 좋아!"
result = classifier(text)
print(result)
# Output: [{'label': '기쁨', 'score': 0.95}]

Model Architecture

• Base Model: klue/roberta-base
• Task: Sequence Classification
• Number of Labels: N/A

Citation

If you use this model, please cite:

@misc{mindcast-model,
  author = {Mindcast Team},
  title = {Mindcast Emotion Classifier},
  year = {2025},
  publisher = {HuggingFace},
  howpublished = {\url{https://huggingface.co/merrybabyxmas/mindcast-emotion-sc-only}},
}

Contact

For questions or feedback, please open an issue on the model repository.

---

This model card was automatically generated.