A Pointer Generator with a BERT encoder
As a toy set, we consider DUC2003 summarization dataset as train, DUC2004 summarization dataset as development and Gigaword data as prediction dataset.
The train and dev dataset are pre-processed into text files with each line containing tab separated input and summaries.
Note: For the model to run, preprocess your training data into this format.
Data Link- https://github.com/harvardnlp/sent-summary
Install requirements by-
pip install -r requirements.txt
- See, Abigail, Peter J. Liu, and Christopher D. Manning. "Get to the point: Summarization with pointer-generator networks." arXiv preprint arXiv:1704.04368 (2017).
- Devlin, Jacob, et al. "Bert: Pre-training of deep bidirectional transformers for language understanding." arXiv preprint arXiv:1810.04805 (2018).
Not ready to use