Code for EMNLP 2023 paper "Enhancing Neural Machine Translation with Semantic Units"
Please refer to https://github.com/shrango/Words-Pair-Encoding for guidelines.
We put example training and generating scripts in the scripts file. Following the instructions in the scripts to fill the path of processed data. Although the scripts are for En-Ro task, you can change the lang_ids as your wish.
@inproceedings{
huang2023enhancing,
title={Enhancing Neural Machine Translation with Semantic Units},
author={Huang, Langlin and Gu, Shuhao and Zhang, Zhuocheng and Feng, Yang
},
booktitle={Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing},
year={2023},
}