spencerbraun / sophisticated_style_transfer

Code for paper "Generative Text Style Transfer for Improved Language Sophistication"

Home Page:http://cs230.stanford.edu/projects_winter_2020/reports/32069807.pdf

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

sophisticated_style_transfer

This is the main code repository for the paper Generative Text Style Transfer for Improved Language Sophistication.

For code corresponding to the Style Transfer model, see https://github.com/spencerbraun/style-transformer, a forked project from https://github.com/fastnlp/style-transformer

For code corresponding to the ARAE model, see https://github.com/spencerbraun/ARAE, a forked project from https://github.com/jakezhaojb/ARAE

Data Processing

All data processing manipulations are performed in the "Data Processing.ipynb" file. Raw ASAP data can be found in data/raw, while the processed data used to train the models can be found in data/processed. All other data is generated in the data processing notebook.

Evaluation

The evaluation metrics found in the paper along with others tested can be found in Evaluation.ipynb

Research

The research folder contains notes on relevant literature, evaluation metrics, as well as experiments run with different autoencoder models.

About

Code for paper "Generative Text Style Transfer for Improved Language Sophistication"

http://cs230.stanford.edu/projects_winter_2020/reports/32069807.pdf

License:MIT License


Languages

Language:Roff 74.4%Language:Jupyter Notebook 17.1%Language:Python 8.3%Language:Shell 0.1%