mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-06-30 08:10:20 +00:00
2ceda3f662
* ggml-cpu: use LUT for converting e8->f32 scales on x86 * add dispatch based on macro