---
license: cc-by-nc-4.0
language:
  - gsw
  - multilingual
inference: false
---

The [SwissBERT](https://huggingface.co/ZurichNLP/swissbert) model ([Vamvas et al., SwissText 2023](https://aclanthology.org/2023.swisstext-1.6/)) extended by a Swiss German adapter that was trained on the character level.

**Note:** This model is experimental and can only be run with our codebase at https://github.com/ZurichNLP/swiss-german-text-encoders, since it uses a custom model architecture.

## Training Data
For continued pre-training, we used the following two datasets of written Swiss German:
1. [SwissCrawl](https://icosys.ch/swisscrawl) ([Linder et al., LREC 2020](https://aclanthology.org/2020.lrec-1.329)), a collection of Swiss German web text (forum discussions, social media).
2. A custom dataset of Swiss German tweets

In addition, we trained the model on an equal amount of Standard German data. We used news articles retrieved from [Swissdox@LiRI](https://t.uzh.ch/1hI).

## License
Attribution-NonCommercial 4.0 International (CC BY-NC 4.0).

## Citation
```bibtex
@inproceedings{vamvas-etal-2024-modular,
      title={Modular Adaptation of Multilingual Encoders to Written Swiss German Dialect},
      author={Jannis Vamvas and No{\"e}mi Aepli and Rico Sennrich},
      booktitle={First Workshop on Modular and Open Multilingual NLP},
      year={2024},
}
```