Re:Turing-NLG: A 17B-parameter language model
by riku_iki on Monday February 10, 2020 @03:18PM
(#22292801)
unfortunately they abstained from participation in more popular SQuAD and Glue benchmarks..
[ Parent
| on HN ]
Re:Turing-NLG: A 17B-parameter language model
by octbash on Monday February 10, 2020 @06:30PM
(#22294535)
Those are question-answering and language-understanding benchmarks respectively, neither of which has been suitable for language generation mode evaluation since GPT-1 was roundly beating by BERT. GPT-2 didn't evaluate on them either.