A Nemo 12B version?
#7
by
danihend
- opened
I used Nemo for the first time yesterday and was impressed. I thought the R1 treatment would be very interesting and now I saw this π
Any plans to do Nemo?
I second this ! ; lots of Nemos would greatly benefit from the R1/Distill treatment.
I have built some Llamas with Deepseek's Distills "tech" - reg, stacker, Brainstorm and MOEs -- would love to see what Nemos could do with reasoning.
Maybe even older "7B" mistral too?