Text Summarization

SentenceTransformers can be used for (extractive) text summarization: The document is broken down into sentences and embedded by SentenceTransformers. Then, we can compute the cosine similarity across all possible sentence combinations.

We then use LexRank to find the most central sentences in the document. These central sentences form a good basis for a summarization of the document.

An example is shown in text_summarization.py