Bert Base 15lang Cased

🚀 bert-base-15lang-cased

We're sharing smaller versions of bert-base-multilingual-cased that support a custom set of languages. These versions maintain the same accuracy as the original model.

📦 Installation

No specific installation steps are provided in the original document.

✨ Features

• Language Support: Handles 15 languages including en, fr, es, de, zh, ar, ru, vi, el, bg, th, tr, hi, ur, and sw.
• Accuracy Preservation: Unlike distilbert-base-multilingual-cased, it gives the same representations as the original model, preserving accuracy.
• Efficiency: Smaller in size and faster to load compared to the original bert-base-multilingual-cased.

📊 Model Comparison

The measurements below were computed on a Google Cloud n1-standard-1 machine (1 vCPU, 3.75 GB):

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("Geotrend/bert-base-15lang-cased")
model = AutoModel.from_pretrained("Geotrend/bert-base-15lang-cased")

📚 Documentation

For more information, please visit our paper: Load What You Need: Smaller Versions of Multilingual BERT.

To generate other smaller versions of multilingual transformers, please visit our Github repo.

📄 License

This project is licensed under the Apache-2.0 license.

📝 How to Cite

@inproceedings{smallermbert,
  title={Load What You Need: Smaller Versions of Multilingual BERT},
  author={Abdaoui, Amine and Pradel, Camille and Sigel, Grégoire},
  booktitle={SustaiNLP / EMNLP},
  year={2020}
}

📞 Contact

Please contact amine@geotrend.fr for any questions, feedback, or requests.