HHousen/TransformerSum
A library for training and evaluating transformer-based models on automatic text summarization tasks.
Not currently ranked — collecting fresh signals.
star history
TransformerSum provides implementations of both extractive and abstractive summarization models using pre-trained transformer architectures. It integrates with Hugging Face Transformers for easy access to various model architectures and uses PyTorch Lightning to handle training loops. The library also includes utilities for converting abstractive summarization datasets into extractive format, enabling comparative research between the two approaches.
Frequently asked
- What is HHousen/TransformerSum?
- A library for training and evaluating transformer-based models on automatic text summarization tasks.
- Is TransformerSum open source?
- Yes — HHousen/TransformerSum is open source, released under the GPL-3.0 license.
- What language is TransformerSum written in?
- HHousen/TransformerSum is primarily written in Python.
- How popular is TransformerSum?
- HHousen/TransformerSum has 440 stars on GitHub.
- Where can I find TransformerSum?
- HHousen/TransformerSum is on GitHub at https://github.com/HHousen/TransformerSum.