chrisliu298/arxiv_ai_gpt2

The GPT-2 (774M) model is capable of generating abstracts given paper titles . It was trained using all research paper titles and abstracts under artificial intelligence (AI), machine learning (LG), computation and language (CL), and computer vision and pattern recognition (CV) on arXiv.org . The original arxiv Archive dataset contains a full archive of metadata about papers on arxv.org, from the start of the site in 1993 to the end of 2019 . An example of a paper’s title and its abstract is shown below . The resulting large model's perplexity score on the test set is 14.9413.95 (14.94,