simple-and-effective-paraphrastic-similarity

Code to train models from "Simple and Effective Paraphrastic Similarity from Parallel Translations".

The code is written in Python 3.7 and requires numpy, scipy, sentencepiece, sacremoses, and PyTorch >= 1.0 libraries.

To get started, download the data files used for training from http://www.cs.cmu.edu/~jwieting and download the STS evaluation data:

wget http://www.cs.cmu.edu/~jwieting/acl19-simple.zip
unzip acl19-simple.zip
rm acl19-simple.zip
bash download_sts17.sh

If you use our code, models, or data for your work please cite:

@inproceedings{wieting19simple,
    title={Simple and Effective Paraphrastic Similarity from Parallel Translations},
    author={Wieting, John and Gimpel, Kevin and Neubig, Graham and Berg-Kirkpatrick, Taylor},
    booktitle={Proceedings of the Association for Computational Linguistics},
    url={https://arxiv.org/abs/1909.13872},
    year={2019}
}

To train sp-average models in language xx on GPU, choices are es, ar, or tr:

python main.py --data-file acl19-simple/en-xx.os.1m.tok.20k.txt --model avg --dim 300 --epochs 10 --share-vocab 1 --dropout 0.3 --sp-model acl19-simple/en-xx.os.1m.tok.sp.20k.model

To train trigram-average models in language xx on GPU:

python main.py --data-file acl19-simple/en-xx.os.1m.tok.txt --model avg --dim 300 --epochs 10 --ngrams 3 --share-vocab 1 --dropout 0.3

To train the bidirectional LSTM models in language xx on GPU:

python main.py --data-file acl19-simple/en-xx.os.1m.tok.20k.txt --model lstm --dim 300 --epochs 10 --dropout 0.3 --scramble-rate 0.3 --share-vocab 1 --share-encoder 1 --sp-model acl19-simple/en-xx.os.1m.tok.sp.20k.model

jwieting/simple-and-effective-paraphrastic-similarity

simple-and-effective-paraphrastic-similarity