Berkeley Document Summarizer


The Berkeley Document Summarizer is a single-document summarization system described in the following paper:

Learning-Based Single-Document Summarization with Compression and Anaphoricity Constraints
Greg Durrett, Taylor Berg-Kirkpatrick, and Dan Klein.
ACL 2016

Project page on GitHub.


Training documents (1 GB)
Training summaries (47 MB)
Test documents (90 MB)
Test summaries (4 MB)

Models for summarization and preprocessing (36 MB)