Strong paraphrase generation using universal transformer - 2020

Software implementation of research in the framework of graduate qualification work

Research of the generation of strong paraphrases and new metrics for the validation of strong paraphrases.

Project structure

Top-level directory structure with a short description of the contents.

.
├── paraphrase          
    |   ├── russian_model   # Scripts for creating and training the neural network Transformer architecture model for the Russian language
    |   └── english_model   # and for English (various training parameters, text corpora and tokenizer)
├── topic_model             # Scripts for creating and learning topic models
├── analys                  # Scripts for calculating various characteristics of text corpora
├── dataset                 # Scripts for collecting and manipulating data sets
├── download                # Raw files new dataset
├── processed               # Files with calculated characteristics of the text corpora and the new generated files
├── report                  # The files are generated reports (including visualizations of topic models)
├── log                     # Logs the process of downloading the new dataset
├── lib                     # Files of the Mallet library for creating latent dirichlet allocation
├── util                    # Util scripts
├── raw_data                # Files text corpora from old research
├── LICENSE                
└── README.md

Installation

Project dependencies:

pandas                   1.0.1
nltk                     3.4.5 +
numpy                    1.16.3 +
tensorboard              1.15.0 +
tensorflow-estimator     1.15.1 +
tensorflow-gpu           1.15.0 +
scikit-learn             0.23.1
scipy                    1.4.1 +
sentencepiece            0.1.86 +
Keras-Applications       1.0.8
Keras-Preprocessing      1.1.0 +
Werkzeug                 1.0.1
astor                    0.8.1
cachetools               4.1.0
chardet                  3.0.4
expressions              0.2.3
hparams                  0.3.0
idna                     2.8 +
joblib                   0.14.1 +
lxml                     4.5.0 +
mock                     4.0.2
opt-einsum               3.2.1
protobuf                 3.11.3
pyasn1                   0.4.8
python-dateutil          2.8.1
pytz                     2019.3 +
regex                    2017.4.5 +
requests                 2.21.0 +
tqdm                     4.31.1 +
typeguard                2.7.1
urllib3                  1.24.3 +

Meta

Kirill Vakhrushev – VK – kirill.vakhrushev@gmail.com

Distributed under the MIT license. See LICENSE for more information.

https://github.com/krikyn/

Contributing

1. Fork it (https://github.com/krikyn/Strong-Paraphrase-Generation-2020/fork)
2. Create your feature branch (git checkout -b feature/fooBar)
3. Commit your changes (git commit -am 'Add some fooBar')
4. Push to the branch (git push origin feature/fooBar)
5. Create a new Pull Request