New and improved Q1_S quants

#1
by LapinMalin - opened

Hi Dan,

Is there any chance you could try to re-quantize this model with the new Q1_S algorithm in llama.cpp?

https://github.com/ggerganov/llama.cpp/pull/5999

Alternatively, could you maybe upload the imatrix.dat file you used for this quantization?

Thanks. :)

I'll see if I can do that. @LapinMalin there you go, enjoy!

I'll see if I can do that. @LapinMalin there you go, enjoy!

Amazing, thank you so much!

Sign up or log in to comment