mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-09-01 01:51:19 +02:00
69c487f4ed
* CUDA: MMQ code deduplication + iquant support * 1 less parallel job for CI build