diff --git a/systems/x86_64-linux/halo/llama-server-27B-MTP.nix b/systems/x86_64-linux/halo/llama-server-27B-MTP.nix index cd58085..d32041c 100644 --- a/systems/x86_64-linux/halo/llama-server-27B-MTP.nix +++ b/systems/x86_64-linux/halo/llama-server-27B-MTP.nix @@ -46,7 +46,8 @@ "-c 262144" "--fit on" "--slot-save-path %C/llama-server/kv-slots-27B-MTP" - "--spec-type draft-mtp --spec-draft-n-max 2" + "--spec-type draft-mtp --spec-draft-n-max 6" + #"--spec-type ngram-mod --spec-ngram-mod-n-match 24 --spec-ngram-mod-n-min 48 --spec-ngram-mod-n-max 64" ]; Restart = "on-failure"; RestartSec = 10;