Could there be 8b, 7b and 1.5b version
#206
by
BenjaminChu
- opened
I hope perlexity can release smaller models, so I could run on local.
BenjaminChu
changed discussion title from
R1-1989
to Could there be 8b, 7b and 1.5b version
That’s going to be super helpful too.
id second that, for us serfs who have limited resources on prem. ( 14 too, that is my 'sweet spot' )
You can check out our similar 8b model here, https://huggingface.co/BlueBeck/LlamaAligned-DeepSeekR1-Distill-8b , it simply aims to retain the reasoning ability of the 8b distil R1 model and to align with Llama 3.3, for better or for worse, rather than making any particular political statement.