If most of the performance gains are hidden in a few middle layers, you can save... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

tribal808 73 days ago | parent | context | favorite | on: Is One Layer Enough? A Single Transformer Layer Ma...

If most of the performance gains are hidden in a few middle layers, you can save a massive amount of compute by freezing the rest

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact