llama : only use default buffer types for the KV cache#10358

slaren merged 1 commit intomasterggml-org/llama.cpp:masterfrom

sl/fix-kv-buftggml-org/llama.cpp:sl/fix-kv-buftCopy head branch name to clipboard