91 points de_sty 3 hours ago 22 comments
robertkarl 2 hours ago | parent
xlayn 2 hours ago | parent
"18 days on 144 H200 GPUs" insane.. didn't thought so much hardware would be needed to train Qwen.
drdaeman 1 hour ago | parent
smallerize 53 minutes ago | parent
drdaeman 48 minutes ago | parent
XCSme 2 hours ago | parent
octoberfranklin 2 hours ago | parent
You might be able to shave that down using a different flavor of H200 (like PCIe) but it would take more hours, so benefits are unclear.
lumost 1 hour ago | parent
octoberfranklin 1 hour ago | parent
But seriously. If there were a cheaper and reliable way, I'm sure the authors would have mentioned it instead of H200 SXM modules. They're trying to brag about cheap training.
Maybe you wanted to use an opportunity to shill tranium.
phs318u 45 minutes ago | parent
This was unnecessary and against HN guidelines to engage in good faith.
woah 1 hour ago | parent
charcircuit 21 minutes ago | parent
It already happened. https://github.com/karpathy/autoresearch
jacquesm 1 hour ago | parent
nblintao 1 hour ago | parent
jacquesm 56 minutes ago | parent
reissbaker 48 minutes ago | parent
jacquesm 41 minutes ago | parent
And no, that won't get you to AGI, that is a completely different goal.
banghua13 21 minutes ago | parent
jacquesm 13 minutes ago | parent
pkoiralap 7 minutes ago | parent
https://huggingface.co/sabicap/sabi-noema
Also HF model here: https://huggingface.co/RadixArk/DaisyRose-27B