From 2426ca563a834d2b0d8a94ba19b2e26d25625608 Mon Sep 17 00:00:00 2001
From: Frank <frank@dankware>
Date: Fri, 7 Aug 2026 16:25:49 -0700
Subject: [PATCH] Add ngram-mod to spec pipeline: -1.2 tok/s regression

- DFlash+ngram-mod: 17.11 tok/s vs DFlash alone: 18.34 tok/s
- ngram overhead not offset by acceptance rate at current settings
- Keeping in place for now, may tune ngram-mod params later
---
 scripts/llama-server.service     | 2 +-
 scripts/serve-llamacpp-dflash.sh | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/scripts/llama-server.service b/scripts/llama-server.service
index 2a76b96..f2a8430 100644
--- a/scripts/llama-server.service
+++ b/scripts/llama-server.service
@@ -6,7 +6,7 @@ After=network.target
 Type=simple
 User=dank
 Group=dank
-ExecStart=/home/dank/llama.cpp/build/bin/llama-server --hf-repo unsloth/Qwen3.6-27B-MTP-GGUF:Q8_0 --spec-draft-hf Alittlehammmer/Qwen3.6-27B-DFlash-GGUF-llama.cpp:Q8_0 --spec-type draft-dflash --spec-draft-n-max 16 --device ROCm0 --spec-draft-device ROCm0 --gpu-layers all --fit on --fit-target 512 --parallel 3 --ctx-size 131072 --batch-size 4096 --port 8080 --host 0.0.0.0 --threads 16 --load-mode none
+ExecStart=/home/dank/llama.cpp/build/bin/llama-server --hf-repo unsloth/Qwen3.6-27B-MTP-GGUF:Q8_0 --spec-draft-hf Alittlehammmer/Qwen3.6-27B-DFlash-GGUF-llama.cpp:Q8_0 --spec-type draft-dflash,ngram-mod --spec-draft-n-max 16 --device ROCm0 --spec-draft-device ROCm0 --gpu-layers all --fit on --fit-target 512 --parallel 3 --ctx-size 131072 --batch-size 4096 --port 8080 --host 0.0.0.0 --threads 16 --load-mode none
 Restart=on-failure
 RestartSec=10
 
diff --git a/scripts/serve-llamacpp-dflash.sh b/scripts/serve-llamacpp-dflash.sh
index 70b33bd..8694d36 100755
--- a/scripts/serve-llamacpp-dflash.sh
+++ b/scripts/serve-llamacpp-dflash.sh
@@ -26,7 +26,7 @@ echo ""
 exec "${SERVER_BIN}" \
     --hf-repo "${TARGET_REPO}" \
     --spec-draft-hf "${DRAFT_REPO}" \
-    --spec-type draft-dflash \
+    --spec-type draft-dflash,ngram-mod \
     --spec-draft-n-max "${DRAFT_N_MAX}" \
     --device ROCm0 \
     --spec-draft-device ROCm0 \