llama.cpp optimal command?

#2
by datayoda - opened

any tips on temp, min-p etc...

The original model page lists recommended inference parameters: https://huggingface.co/MiniMaxAI/MiniMax-M2.7#inference-parameters

We recommend using the following parameters for best performance: temperature=1.0, top_p = 0.95, top_k = 40

Sign up or log in to comment