mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-06-28 07:10:21 +00:00
35266573b9
* implement soft_max * Fix soft_max data race * Temporary fix, wait on each submit