README.md

---
license: other
license_name: license
license_link: LICENSE
base_model:
- google/gemma-2-2b
pipeline_tag: translation
---

# Model Card for GemmaX2-28

## Model Details

### Model Description

GemmaX2-28-2B-Pretrain is a language model that results from continual pretraining of Gemma2-2B on a mix of 56 billion tokens of monolingual and parallel data in 28 different languages — Arabic, Bengali, Czech, German, English, Spanish, Persian, French, Hebrew, Hindi, Indonesian, Italian, Japanese, Khmer, Korean, Lao, Malay, Burmese, Dutch, polish, Portuguese, Russian, Thai, Tagalog, Turkish, Urdu, Vietnamese, Chinese. 

- **Developed by:** Xiaomi
- **Model type:** A 2B parameter model base on Gemma2-2B, we obtained GemmaX2-28-2B-Pretrain by continuing pre-training on a large amount of monolingual and parallel data.
- **Languages:** Arabic, Bengali, Czech, German, English, Spanish, Persian, French, Hebrew, Hindi, Indonesian, Italian, Japanese, Khmer, Korean, Lao, Malay, Burmese, Dutch, polish, Portuguese, Russian, Thai, Tagalog, Turkish, Urdu, Vietnamese, Chinese. 
- **License:**  gemma

### Model Source

- paper: [Multilingual Machine Translation with Open Large Language Models at Practical Scale: An Empirical Study](https://arxiv.org/pdf/2502.02481)

### Model Performance

![Experimental Result](main.png)


### Training Data

We collected monolingual data from [CulturaX](https://huggingface.co/datasets/uonlp/CulturaX) and [MADLAD-400](https://huggingface.co/datasets/allenai/MADLAD-400). For parallel data, we collected all Chinese-centric and English-centric parallel dataset from the [OPUS](https://opus.nlpl.eu/) collection up to Auguest 2024 and underwent a series of filtering processes, such as language detection, semantic duplication filtering, quality filtering, and more.

## Citation 

```bibtex
@misc{cui2025multilingualmachinetranslationopen,
      title={Multilingual Machine Translation with Open Large Language Models at Practical Scale: An Empirical Study}, 
      author={Menglong Cui and Pengzhi Gao and Wei Liu and Jian Luan and Bin Wang},
      year={2025},
      eprint={2502.02481},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2502.02481}, 
}
```