NewsSum
A neural news summarization tool. Collect google news topics, crawl related news articles, generate summaries. Including the following units:
- BasicSum: traditional freq-based news-summary generator.
- LSTM-Attetion: Neural network model that summarizes news.
- WikiNews Dataset: A list of news articles crawled from WikiNews, every element(event) of the list contains at least three articles from mainstream website talking about the event.
Run
python3 main.py
Demo
Pretrained Model Download
https://drive.google.com/file/d/0BwzJqD-PyhaoTVhZZVR2UTN6TU0/view?usp=sharing
Data
https://docs.google.com/uc?id=0B0Obe9L1qtsnSXZEd0JCenIyejg&export=download
Dependency
NLTK
numpy
Tensorflow
Newspaper
Acknowledgement
NeuralSum For their dataset and dataset reader and LSTM architecture.
License
MIT