Default setting of llama.cpp and settings required for go-llama

[hiqsociety]

Getting oomed. not sure what other settings i need to do:

llama.cpp that works: ./main -m models/speechless-llama2-13b.Q3_K_S.gguf -n 3400 -c 1920 -p "write a blog about llama:" -e --temp 0.3 --mirostat 2

oomed reference: https://github.com/go-skynet/go-llama.cpp/discussions/237

sorry asking here because waited for days without reply.

[hiqsociety]

@mudler @deadprogram @mauromorales @jrc2139 @soleblaze any help here? sorry been waiting for days over this.