mirror of
https://github.com/ggml-org/llama.cpp.git
synced 2026-08-25 22:21:03 +02:00
af0a5b6163
* server: normalize token probabilities * fix temperature == 0.0f