mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2025-11-02 09:12:03 +00:00
* YaRN : store rope scaling type as int32_t in memory * llama : store mapped names as const char *
40 KiB
40 KiB