llama: GPU-resident LRU cache for host-offloaded MoE expert weights - #27861

llama: GPU-resident LRU cache for host-offloaded MoE expert weights#27861

csantiago78 wants to merge 1 commit into

ggml-org:masterggml-org/llama.cpp:masterfrom

csantiago78:moe-expert-cachecsantiago78/llama.cpp:moe-expert-cacheCopy head branch name to clipboard