Pubmedbert Bio Ext Summ

🚀 PubMedBERT for biomedical extractive summarization

This project is about fine - tuning PubMedBERT on MS² for biomedical extractive summarization, aiming to provide effective document summarization solutions in the biomedical field.

🚀 Quick Start

This is a project related to the work of the author's Bachelor's thesis. It fine - tunes PubMedBERT on MS² for extractive summarization. The model architecture is similar to BERTSum. The training code can be found at biomed-ext-summ.

💻 Usage Examples

Basic Usage

summarizer = pipeline("summarization",
  model = "NotXia/pubmedbert-bio-ext-summ",
  tokenizer = AutoTokenizer.from_pretrained("NotXia/pubmedbert-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 several strategies to summarize the document:

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

📄 License

This project is licensed under the Apache-2.0 license.