nvidia/Mistral-NeMo-12B-Base
#156
by
DazzlingXeno
- opened
Could you do a quant of that please?
nvidia's version seems to be a foreign format of the regular version. I don't think this can / has to be requested.
yes, it's in their proprietary format, thanks for pointing this out and saving me work @Henk717 :)
mradermacher
changed discussion status to
closed
Ahhh damn, no worries. If therw is a base version of the 12b I'd be eager to use it. Great work you do mradermacher btw. Much appreciated by many I am sure.
damn nvidia pushing their proprietary shit...
Don't accuse me of advertising, but meta-llama/Meta-Llama-3-70B-Instruct is trustworthy.
12b? 70b? I trust it to be too big to replace a 12b :)
meta-llama/Meta-Llama-3-8B-Instruct :)
The models from Nvidia are just average, but their APIs are really user-friendly. :)
then i have no regrets :=)