Machine Translation (MT) Models

This repository hosts two machine translation models that translate Maltese sentences to English.

Models

Model 1:

This model is sourced from Kurt Abela's MSc Thesis repository. If you utilize this model, kindly acknowledge it by citing the following:

@mastersthesis{kurt-abela-thesis,
  title={Machine Translation System for Low Resource Languages},
  author={Kurt Abela},
  school={University of Malta},
  address={Malta},
  year={2023},
}

Model 2:

The second model is the machine translation model used for the iWSLT 2023 shared task submission. For citation purposes, please use:

@inproceedings{williams2023dfki,
  title={UM-DFKI Maltese Speech Translation},
  author={Williams, Aiden and Abela, Kurt and Kumar, Rishu and B{\"a}r, Martin and Billinghurst, Hannah and Micallef, Kurt and Samin, Ahnaf Mozib and DeMarco, Andrea and van der Plas, Lonneke and Borg, Claudia},
  booktitle={Proceedings of the 20th International Conference on Spoken Language Translation (IWSLT 2023)},
  pages={433--441},
  year={2023}
}

How to Use

All models are Fairseq models. Each directory contains the best checkpoint (in the ./checkpoints/mt_en directory). The dictionaries are kept in the ./data_out/mt_en directory. The sentencepiece model and vocabulary can be found in the ./models/mt_en directory.

To use the models, we guide users to the Fairseq documentation.

Contact

For any inquiries or issues regarding the models, feel free to contact us.