logoalt Hacker News

m_w_yesterday at 5:03 PM0 repliesview on HN

Very impressive headline benchmark numbers. I expected a step change, but not past Fable. That said - it all depends on whether the classifiers make the model unusable...