LlamaCpp change breaks Q4_0, Q4_1 and Q8_0 models
su77ungr opened this issue · 0 comments
su77ungr commented
Issue you'd like to raise.
Not really needed with our default q5 but it should be noted and added so it's within the doc
ggerganov/llama.cpp#1508 (comment)
just a note to myself
Suggestion:
No response