Longformer Bio Ext Summ

🚀 Longformer for biomedical extractive summarization

This project is a fine - tuned Longformer model for biomedical extractive summarization, which is the result of the author's Bachelor's thesis work.

🚀 Quick Start

The model is a Longformer fine - tuned on MS^2 for extractive summarization. The work is related to the author's Bachelor's thesis. The model architecture is similar to BERTSum, and the training code is available at biomed-ext-summ.

💻 Usage Examples

Basic Usage

summarizer = pipeline("summarization",
  model = "NotXia/longformer-bio-ext-summ",
  tokenizer = AutoTokenizer.from_pretrained("NotXia/longformer-bio-ext-summ"),
  trust_remote_code = True,
  device = 0
)

sentences = ["sent1.", "sent2.", "sent3?"]
summarizer({"sentences": sentences}, strategy="count", strategy_args=2)
>>> (['sent1.', 'sent2.'], [0, 1])

Advanced Usage

There are different strategies to summarize the document:

• length: Generate a summary with a maximum length, where strategy_args represents the maximum length.
• count: Generate a summary with the given number of sentences, where strategy_args is the number of sentences.
• ratio: Generate a summary proportional to the length of the document, where strategy_args is the ratio in the range [0, 1].
• threshold: Generate a summary only with sentences having a score higher than a given value, where strategy_args is the minimum score.

📄 License

The project is licensed under the Apache - 2.0 license.