Skip to content

Comment on Training a 3.8B LLM to 0.384 CORE for $998parent

Comments

I've been training different architecture 1.33b and 3.33b models using nvidia megatron for about a month. You can train a 1.33b on a 46gb mac and 3.33b is about as big as you can go with a single 96gb blackwell with regular checkpoints etc

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.