TheBloke

/

orca_mini_7B-GGML

like 21

I get the following error when trying to load on M2 metal with llama.cpp

Asserting on type 8
GGML_ASSERT: ggml-metal.m:706: false && "not implemented"
GGML_ASSERT: ggml-metal.m:758: false && "not implemented"
Abort trap: 6

Model works find when running llama.cpp without ngl flag

Thanks for all your work