最終更新:2025-08-19 (火) 00:17:34 (91d)  

llama.cpp/設定
Top / llama.cpp / 設定

--n-cpu-moe

  • keep the Mixture of Experts (MoE) weights of the first N layers in the CPU

--cpu-moe?

  • keep all Mixture of Experts (MoE) weights in the CPU