TextSummarization

Text Summarization in Pytorch

The aim of this repo is to explore all the different ways to perform text summarization.

Dataset

The dataset used here is Amazon Reviews.

Current Implemnted Models

• Sequence to Sequence with attention

Future Models

• Sequence to Sequence with attention with RNN weights pre-trained from WikiText-103. Weights can be found here
• AWD-LSTM for ASGD Weight-Dropped LSTM
• Pointer-Generator Networks with Coverage
• Pointer-Generator Networks with Coverage with RNN weights pre-trained from WikiText-103. Weights can be found here

This is still a working repo. Stay tuned...

KeyWords

• pytorch
• spacy
• multi-processing
• cyclical learning rates