mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-06-30 00:00:23 +00:00
69c487f4ed
* CUDA: MMQ code deduplication + iquant support * 1 less parallel job for CI build