feat: frank-default-large → llama.cpp Qwen3.6-27B Q8_K_P + DFlash draft + ngram-mod (bench pending)

scripts/serve-llama-frank-large.sh

@@ -0,0 +1,74 @@
+#!/usr/bin/env bash
+# Serve HauhauCS Qwen3.6-27B-Uncensored-HauhauCS-Aggressive:Q8_K_P
+# with DFlash draft + ngram-mod speculative decoding on dankclaw
+# (AMD Strix Halo gfx1151, ROCm or Vulkan backend).
+# Runtime: llama.cpp llama-server.
+#
+# Tunables come from frank-default-large.env in this directory unless the
+# FRANK_LARGE_ENV_FILE environment variable overrides that path.
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+ENV_FILE="${FRANK_LARGE_ENV_FILE:-$SCRIPT_DIR/frank-default-large.env}"
+if [[ -f "$ENV_FILE" ]]; then
+    set -a
+    # shellcheck disable=SC1090
+    source "$ENV_FILE"
+    set +a
+fi
+
+# --- defaults (used when the env file does not set them) -----------------
+: "${MODEL_DIR:=/home/dank/models}"
+: "${MODEL_FILE:=Qwen3.6-27B-Uncensored-HauhauCS-Aggressive-Q8_K_P.gguf}"
+: "${DRAFT_FILE:=Qwen3.6-27B-DFlash-Q8_0.gguf}"
+: "${ALIAS:=frank-default-large}"
+: "${HOST:=0.0.0.0}"
+: "${PORT:=7770}"
+: "${CTX_SIZE:=393216}"
+: "${PARALLEL:=6}"
+: "${SPEC_TYPE:=draft-dflash,ngram-mod}"
+: "${SPEC_DRAFT_N_MAX:=8}"
+: "${SPEC_NGRAM_MOD_N_MIN:=48}"
+: "${SPEC_NGRAM_MOD_N_MAX:=64}"
+: "${SPEC_NGRAM_MOD_N_MATCH:=24}"
+: "${DEVICE:=ROCm0}"
+: "${N_GPU_LAYERS:=all}"
+: "${CACHE_TYPE_K:=q8_0}"
+: "${CACHE_TYPE_V:=f16}"
+: "${BATCH_SIZE:=2048}"
+: "${UBATCH_SIZE:=512}"
+: "${THREADS:=8}"
+: "${THREADS_HTTP:=8}"
+
+LLAMA_SERVER="${LLAMA_SERVER:-/home/dank/llama.cpp/build/bin/llama-server}"
+MODEL="$MODEL_DIR/$MODEL_FILE"
+DRAFT="$MODEL_DIR/$DRAFT_FILE"
+
+for f in "$MODEL" "$DRAFT"; do
+    if [[ ! -f "$f" ]]; then
+        echo "missing required file: $f" >&2
+        exit 1
+    fi
+done
+
+exec "$LLAMA_SERVER" \
+    --model "$MODEL" \
+    --spec-draft-model "$DRAFT" \
+    --spec-type "$SPEC_TYPE" \
+    --spec-draft-n-max "$SPEC_DRAFT_N_MAX" \
+    --spec-ngram-mod-n-min "$SPEC_NGRAM_MOD_N_MIN" \
+    --spec-ngram-mod-n-max "$SPEC_NGRAM_MOD_N_MAX" \
+    --spec-ngram-mod-n-match "$SPEC_NGRAM_MOD_N_MATCH" \
+    --device "$DEVICE" \
+    --gpu-layers "$N_GPU_LAYERS" \
+    --cache-type-k "$CACHE_TYPE_K" \
+    --cache-type-v "$CACHE_TYPE_V" \
+    --ctx-size "$CTX_SIZE" \
+    --parallel "$PARALLEL" \
+    --batch-size "$BATCH_SIZE" \
+    --ubatch-size "$UBATCH_SIZE" \
+    --threads "$THREADS" \
+    --threads-http "$THREADS_HTTP" \
+    --alias "$ALIAS" \
+    --host "$HOST" \
+    --port "$PORT"