Is this the first major model to be natively FP8? I was wondering why people had... | Hacker News

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

login

modeless 12 months ago | parent | context | favorite | on: Grok

Is this the first major model to be natively FP8? I was wondering why people hadn't done it yet. Seems like a big win when hardware supports it.

a_wild_dandan 12 months ago [–]

No, e.g. Yi-34B.

modeless 12 months ago | [–]

As far as I can tell Yi-34B is natively 16 bit float, the 8 bit version is quantized. https://huggingface.co/01-ai/Yi-34B#quantization

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