Best Abstractive
Updated Daily
The BART Large CNN is a transformer-based model designed for abstractive text summarization. It utilizes convolutional neural networks to process text and reconstruct corrupted input, creating concise summaries. This model was trained on extensive data including the C4 dataset and is particularly us...
PEGASUS CNN/DailyMail is a large language model developed by Google. It utilizes a transformer architecture and was specifically trained on a massive dataset of CNN and Daily Mail news articles alongside their corresponding summaries. This allows it to produce abstractive text summaries – meaning it...
BRIO CNN/DailyMail is an abstractive text summarization model built upon a neural network architecture. It utilizes data from CNN and the Daily Mail to produce concise summaries of news articles. The model’s strength lies in its ability to retain crucial information and maintain factual accuracy dur...
DistilBART CNN 12-6 is a computationally efficient text summarization model built by distilling BART and incorporating a convolutional neural network for improved factual recall and coherence in generated summaries.
ProphetNet CNN/DailyMail utilizes a novel self-supervised pretraining objective alongside a Transformer architecture to generate abstractive summaries of news articles from the CNN/DailyMail dataset. (168 characters)
