mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-07-02 09:10:21 +00:00
28103f4832
* Server: add tests for consistent results * sampling: separate rng per sampling context