DTMT: A Novel Deep Transition Architecture for Neural Machine Translation

Contents

• Introduction
• Usage
• Requirements
• Citation

Introduction

Implementation of DTMT (https://arxiv.org/abs/1812.07807) with incremental decoding. The implementation is based on [THUMT] (https://github.com/THUNLP-MT/THUMT).

Usage

• Training

sh run_train.sh

• Evaluation and Testing

sh run_test.sh

Requirements

• tensorflow 1.10
• python 2.7

Citation

Please cite the following paper if you use the code:

@inproceedings{meng2019dtmt,
  title={{DTMT:} {A} novel deep transition architecture for neural machine translation},
  author={Meng, Fandong and Zhang, Jinchao},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={33},
  pages={224--231},
  year={2019}
}