Accetable results with i1-Q5_K_M or i1-Q6_0 ?
#1
by wijjjj - opened
Hi, I'm trying to make it work. I would really appreciate if anyone could share their configuration, how they made it work stable.
So far I only had garbage came out of any of the quants, unfortunately also from this repo (used i1-Q5_K_M and i1-Q6_0).
Thanks a lot in advance, appreciated!
+1 - i just get nonsense output with llama.cpp