liangyupu/DoTA_dataset

Document Image Machine Translation with Dynamic Multi-pre-trained Models Assembling

This is the official repository for DoTA dataset (Document image machine Translation dataset of ArXiv articles in markdown format) introduced by the following paper: Document Image Machine Translation with Dynamic Multi-pre-trained Models Assembling (NAACL 2024 Main)

In addition to the 126K samples mentioned in the paper, we provide all 139K samples that have not been filtered. Each sample contains original English image, transcripted English mmd file and translated Chinese/French/German mmd file. Samples used in the paper are listed in a json file.

Text files can be decompressed as follows:

tar -xzvf zh_mmd.tar.gz -C ./

Image files can be decompressed as follows:

cat imgs.tar.gz.* | tar -xzvf - -C ./

If you want to use our dataset, please cite as follows:

@inproceedings{
liang2024document,
title={Document Image Machine Translation with Dynamic Multi-pre-trained Models Assembling},
author={Yupu Liang and Yaping Zhang and Cong MA and Zhiyang Zhang and Yang Zhao and Lu Xiang and Chengqing Zong and Yu Zhou},
booktitle={2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics},
year={2024},
url={https://openreview.net/forum?id=XH2TgKlXWv}
}