Callidior/bert2bert-base-arxiv-titlegen

The model is a BERT2BERT Encoder-Decoder using the official bert-base-uncased checkpoint as initialization for the encoder and decoder . It was fine-tuned on 318,500 computer science papers posted on arXiv.org between 2007 and 2020 . It achieved a 26.3% Rouge2 F1-Score on held-out validation data . The live demo is available at http://paper-titles.eyr.r.com/receive-title-generating-simulation-generated-at-a-paper-title generating title for computer science papers given an abstract . It has a 26,