Hacker News new | ask | show | jobs
by mentalgear 265 days ago
seems like they use the normal transformer architecture versus deep fold's more specialised machine-learning approaches.