SeqGAN VAE Paraphrasing

About The Project

My project is called Paraphrasing. This is an implementation of An End-to-End Generative Architecture for Paraphrase Generation.

Getting Started

To get started, you should have prior knowledge on Python and Pytorch at first. A few resources to get you started if this is your first Python or Tensorflow project:

• Pytorch Tutorials
• Python for Beginners

Installation and Run

1. Clone the repo

   git clone https://github.com/phkhanhtrinh23/seqgan_vae_paraphrasing.git

2. Use any code editor to open the folder seqgan_vae_paraphrasing.

Step-by-step

1. Read and run data.py to convert data/train.csv to a compatible format. The dataset originates from Quora Question Pairs (QQP).

2. Read and run train.py to train the SeqGAN VAE model. The model architecture originates from "An End-to-End Generative Architecture for Paraphrase Generation".

Results

Description:

• Inp: the input data.
• Pre: the prediction from the model.
• Tar: the targe/label data (groundtruth).

Note 1: <eos> is just the end-of-sentence token.

Note 2: As you can witness, QQP just covers paraphrasing on question so this model may not work well on normal sentences. Moreover, some of the QQP's data are not good enough to the model because of the low quality of inputs and labels. Sometimes, our model has much better paraphrases than the QQP's labels.

Contribution

Contributions are what make GitHub such an amazing place to be learn, inspire, and create. Any contributions you make are greatly appreciated.

1. Fork the project
2. Create your Contribute branch: git checkout -b contribute/Contribute
3. Commit your changes: git commit -m 'add your messages'
4. Push to the branch: git push origin contribute/Contribute
5. Open a pull request

Contact

Email: phkhanhtrinh23@gmail.com

Project Link: https://github.com/phkhanhtrinh23/seqgan_vae_paraphrasing.git