Indonesian Regional Languages Identifier

Fine-tuned XLM-RoBERTa model for identifying 11 Indonesian regional languages + English.

Supported Languages

• 🇮🇩 Indonesian (Bahasa Indonesia)
• Acehnese (Bahasa Aceh)
• Balinese (Basa Bali)
• Banjarese (Bahasa Banjar)
• Buginese (Basa Ugi)
• Javanese (Basa Jawa)
• Madurese (Basa Madhura)
• Minangkabau (Baso Minang)
• Ngaju (Basa Ngaju)
• Sundanese (Basa Sunda)
• Toba Batak (Hata Batak Toba)
• 🇬🇧 English

Model Performance

• Accuracy: 0.9783
• F1 Macro: 0.9783
• F1 Weighted: 0.9783
• Precision: 0.9785
• Recall: 0.9783

Usage

from transformers import pipeline

# Load model
classifier = pipeline("text-classification", model="YOUR_USERNAME/xlm-roberta-indonesian-languages")

# Single prediction
result = classifier("Sugeng enjing, piye kabare?")
print(result)
# Output: [{'label': 'javanese', 'score': 0.9876}]

# Batch prediction
texts = [
 "Selamat pagi, apa kabar?",
 "Wilujeng enjing, kumaha damang?",
 "Good morning, how are you?"
]

results = classifier(texts)
for text, result in zip(texts, results):
 print(f"{text} -> {result['label']} ({result['score']:.4f})")

Training Details

• Base Model: xlm-roberta-base
• Training Samples: 6000
• Validation Samples: 1200
• Epochs: 5
• Learning Rate: 2e-05
• Batch Size: 16
• Training Date: 20251124_070409

Citation

If you use this model, please cite:

@misc{indonesian-language-id,
 author = {Raihan Hidayatullah Djunaedi},
 title = {Indonesian Regional Languages Identifier},
 year = {2025},
 publisher = {Hugging Face},
 url = {https://huggingface.co/nahiar/xlm-roberta-indonesian-languages}
}

Base model

FacebookAI/xlm-roberta-base

(4081)

A collection of Indonesian NLP models for various text classification tasks such as spam detection, hate speech, abusive language, and more. Suitable • 8 items • Updated Mar 30