New best story on Hacker News: GPT-J-6B – A 6 billion parameter, autoregressive text generation model

GPT-J-6B – A 6 billion parameter, autoregressive text generation model
567 by canada_dry | 127 comments on Hacker News.