mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-06-30 16:20:20 +00:00
c5c64f72ac
* llama : disable Direct IO by default * cont : override mmap if supported