fp16 saves a *lot* of memory and is worth doing. I've not had trouble fine tunin... | Hacker News

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

		slashcom on Aug 20, 2019 \| parent \| context \| favorite \| on: GPT-2: 6-Month Follow-Up fp16 saves a lot of memory and is worth doing. I've not had trouble fine tuning all these models with fp16.

sdan on Aug 21, 2019 [–]

Have you fine tuned 774 successfully using a single GPU?

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