mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-06-30 00:00:23 +00:00
13b339bcd9
* server : do not default to multiple slots with speculative decoding * cont : fix