diff --git a/scripts/start-server.sh b/scripts/start-server.sh index 7f0d36b..2204ec4 100755 --- a/scripts/start-server.sh +++ b/scripts/start-server.sh @@ -13,6 +13,9 @@ export NCCL_PROTO=SIMPLE exec /opt/llama.cpp/build-nccl/bin/llama-server \ -m /opt/models/gguf/Qwen3.8-27B-Uncensored-HauhauCS-Aggressive-Q4_K_P.gguf \ + --spec-type ngram-simple \ + --spec-ngram-simple-size-n 24 \ + --spec-ngram-simple-size-m 4 \ --mmproj /opt/models/gguf/mmproj-Qwen3.8-27B-Uncensored-f16.gguf \ -ngl 99 \ -c 262144 \