Similarity-aware-NMT provides a general NMT framework to incorporate multiple dimensional translation memory based features as prior knowledge into Neural Machine Translation. Meanwhile, Similarity-aware-NMT preprocesses the testing set to get high-potential testing set, like string high-potential,structural high-potential and temporal high-potential Please refer to the following paper for details:
Tianfu Zhang, Heyan Huang, Chong Feng, Xiaochi Wei. 2020. Similarity-aware neural machine translation: reducing human translator efforts by leveraging high-potential sentences with translation memory. Neural Computing and Applications, pp. 1–13.2020.
Similarity-aware-NMT is built on top of THUMT. It requires THEANO 0.8.2 or above version (0.8.2 is recommended)
pip install theano==0.8.2
The source code is dual licensed. Open source licensing is under the BSD-3-Clause, which allows free use for research purposes.