bart-large-cnn
BART-large fine-tuned on CNN/DailyMail for extractive-abstractive summarization. It was a leading open-source summarization model at release and remains a useful baseline for news article summarization tasks.
1,559,897 ↓ · 1,599 ♡
summarization models
4 models · ranked by HuggingFace downloads
distilbart-cnn-12-6
distilbart-cnn-12-6 performs text summarization by encoding the full source and decoding a compressed representation.
861,252 ↓ · 322 ♡
rut5_base_headline_gen_telegram
rut5_base_headline_gen_telegram condenses source documents into shorter text. It produces fluent summaries rather than extracting existing sentences verbatim.
630,532 ↓ · 9 ♡
bart-large-cnn-samsum
As a bart-based open-weight model, bart-large-cnn-samsum focuses on summarization. The MIT license keeps bart-large-cnn-samsum unrestricted for commercial reuse. Check the bart-large-cnn-samsum model card for benchmarks and intended use before adopting it.
295,142 ↓ · 267 ♡