logoalt Hacker News

bombcaryesterday at 6:54 PM3 repliesview on HN

I have no idea what AI involves, but "training" sounds like a one-and-done - but how is the result "stored"? If you have trained up a Gemini, can you "clone" it and if so, what is needed?

I was under the impression that all these GPUs and such were needed to run the AI, not only ingest the data.


Replies

DougBTXyesterday at 9:06 PM

> but how is the result "stored"

Like this: https://huggingface.co/docs/safetensors/index

esafakyesterday at 7:12 PM

Yes, serving requires infra, too. But you can use infra optimized for serving; nvidia GPUs are not the only game in town.

tefkahyesterday at 7:34 PM

Theoretically it would be much less expensive to just continue to run the existing models, but ofc none of the current leaders are going to stop training new ones any time soon.

show 1 reply