dmmiller612/bert-extractive-summarizer
Extractive text summarization library using BERT embeddings and clustering to identify the most representative sentences.
Velocity · 7d
+0.6
★ / day
Trend
→steady
star history
This library leverages HuggingFace PyTorch transformers to embed sentences, then applies clustering algorithms to find sentences closest to cluster centroids for summarization. It also incorporates coreference resolution using neuralcoref to improve summary coherence by resolving pronouns and references. The tool runs on GPU by default when available.