Skip to content

CS 337 Artificial Intelligence and Machine Learning Project

Notifications You must be signed in to change notification settings

harpoonix/Text-Style-Transfer

Repository files navigation

Text-Style-Transfer

Text style transfer is an important task in natural language generation, which aims to control certain attributes in the generated text, such as politeness, emotion, humor, and many others. It has a long history in the field of natural language processing, and recently has re-gained significant attention thanks to the promising performance brought by deep neural models. In this project, we study two methods and architectures employed for this task, compare their generated results and the results of our attempt at reproducibility. In some places, we have also attempted to analyse how our modifications to these architectures compare to the ones designed by the authors. We end with some insights on how the task of text style transfer can be approached, with possible extensions to our modifications that can make it work better.

Paper by Ke Wang, Hang Hua and Xiojun Wan

Our attempt at reproducibility with some architecture modifications is in this folder.

Paper by Ning Dai, Jianze Liang, Xipeng Qiu, Xuanjing Huang

Our attempt at reproducibility with some architecture modifications is in this folder.

Demo

A demo of the style transfer task using our best trained model can be found in this python script.

Our Saved weights

Report

https://drive.google.com/file/d/1SYTfO6GON6mF3fRq9uBd8rnooW4l5CYs/view?usp=sharing

About

CS 337 Artificial Intelligence and Machine Learning Project

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages