Final-project--document-summarization-GAN

This is a proejct for university college london master thesis.
department of computer science Web Science and big data analytics.

This is a generative adversarial networks used in document summariztaion area. The generator is a sequence to sequence model. Discriminator is a CNN model.

Run the code:

1. run data_preprocessing_and_vocabulary_bulid
2. run summarizationGAN

When runing the summariztaionGAN, it has several steps:

1. pre-train the generator
2. decode  
3. pre train discriminator
4. run GAN
5. test generator model

For using the Jupyter notebook, You need to restart the kernel before each step.