They aren’t, there is a 1.58 version of deepseek that’s like 200gb instead of 70... | Hacker News

Hacker News new | past | comments | ask | show | jobs | submit

login

		yieldcrv 1 day ago \| parent \| context \| favorite \| on: BitNet b1.58 2B4T Technical Report They aren’t, there is a 1.58 version of deepseek that’s like 200gb instead of 700

logicchains 1 day ago [–]

That's not a real BitNet, it's just a post-training quantisation, and its performance suffers compared to if it was trained from scratch at 1.58 bits.

Join us for AI Startup School this June 16-17 in San Francisco!
Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact