Marfin Emotion

🚀 Marfin Emotion Detection Model 🎵

This model is fine-tuned from MoritzLaurer/mDeBERTa-v3-base-mnli-xnli for emotion detection tasks based on chat context. It is specifically optimized for Indonesian and English, aiming to accurately detect emotions within chat conversations.

🚀 Quick Start

The model predicts the relationship between user chat input and emotional hypotheses. It can detect emotions such as senang (happy), sedih (sad), marah (angry), takut (fear), and cinta (love). This is useful for emotion-based music recommendation, sentiment analysis in real - time chat apps, and AI - driven mood detection systems.

✨ Features

• Multilingual Support: Optimized for both Indonesian and English, making it suitable for a wide range of users.
• Zero - Shot Classification: Can perform emotion detection without the need for extensive fine - tuning on specific datasets.

📦 Installation

Since this model uses the transformers library, you can install it using the following command:

pip install transformers

💻 Usage Examples

Basic Usage

from transformers import pipeline

classifier = pipeline("zero-shot-classification", model="MarfinF/marfin_emotion")

text = "Aku lagi sedih banget hari ini"
labels = ["senang", "sedih", "marah", "takut", "cinta"]
result = classifier(text, candidate_labels=labels)

print(result)

Advanced Usage

You can expand the list of candidate labels or use different input texts to perform more complex emotion detection tasks. For example, you can add more nuanced emotion labels or use longer chat transcripts as input.

from transformers import pipeline

classifier = pipeline("zero-shot-classification", model="MarfinF/marfin_emotion")

text = "Aku sangat senang karena berhasil menyelesaikan proyek besar ini!"
labels = ["senang", "sedih", "marah", "takut", "cinta", "puas", "tersinggung"]
result = classifier(text, candidate_labels=labels)

print(result)

🔧 Technical Details

• Base Model: mDeBERTa - v3 - base - mnli - xnli
• Fine - Tuning: Fine - tuned with a custom NLI - style dataset.
• Metrics: The performance of the model is evaluated using Accuracy.

📄 License

This project is licensed under the Apache - 2.0 license.

📚 Documentation

Model Information

Tags

Zero - Shot Classification, Emotion, Mood, Indonesian, English