hiennthp/vietnamese-greenwashing-detector

Vietnamese Greenwashing Detection Model (v1.0.0)

Model Description

This model is a fine-tuned version of vinai/phobert-base for detecting greenwashing claims in Vietnamese ESG (Environmental, Social, Governance) reports from banking institutions.

Version: v1.0.0
Last Updated: 2025-11-25 02:58:11 UTC
Base Model: vinai/phobert-base

Performance Metrics

Metric	Score
Accuracy	0.7500
F1 Score	0.6429
Precision	0.5625
Recall	0.7500

Training Data

• Training Samples: 14
• Validation Samples: 4
• Total Samples: 18

Intended Use

This model is designed to:

• Detect greenwashing claims in Vietnamese ESG reports
• Support ESG analysts in reviewing sustainability disclosures
• Flag potential misleading environmental claims

Usage

from transformers import pipeline

classifier = pipeline(
    "text-classification",
    model="hiennthp/vietnamese-greenwashing-detector",
    tokenizer="hiennthp/vietnamese-greenwashing-detector"
)

text = "Ngân hàng cam kết giảm 30% khí thải carbon trong năm 2024."
result = classifier(text)
print(result)

Limitations

• Trained specifically on Vietnamese banking ESG reports
• May not generalize well to other industries or languages
• Should be used as a supporting tool, not sole decision-maker

Training Procedure

• Base Model: vinai/phobert-base
• Training Framework: Hugging Face Transformers
• Optimizer: AdamW
• Learning Rate: 2e-05
• Epochs: 3
• Batch Size: 4

Citation

If you use this model, please cite:

@misc{vietnamese-greenwashing-detector-v1-0-0,
  author = {hiennthp},
  title = {Vietnamese Greenwashing Detection Model},
  year = {2024},
  publisher = {Hugging Face},
  version = {v1.0.0},
  url = {https://huggingface.co/hiennthp/vietnamese-greenwashing-detector}
}

License

MIT License

Contact

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