mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-09-01 10:01:23 +02:00
56db501e73
* mtmd: use pillow-accurate resize algo, correct resize_algo for all models * speed optimization