mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-07-16 17:35:58 +02:00
13b339bcd9
* server : do not default to multiple slots with speculative decoding * cont : fix