dank/serverAppConfigs
1
0
Fork 0
Code Issues 6 Pull Requests Actions Packages Projects Releases Wiki Activity

Switch frank-default-small to Ornith Q6_K

Browse Source
This commit is contained in:
Frank
2026-08-25 00:29:14 -07:00
parent 3cc99d81cc
commit af6ffeeb64
8 changed files with 115 additions and 60 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/frank-default-small.env
+18 -14
View File
@@ -1,34 +1,37 @@
# Tunables for frank-default-small (Ornith-1.5-35B-A3B CRACK Q8_0, MTP MoE,
# Tunables for frank-default-small (Ornith-1.5-35B-A3B CRACK Q6_K, MTP MoE,
# Dankserver, RTX 3090).
#
# Runtime: upstream llama.cpp llama-server (b10121). RTX 3090 24 GB; Ryzen
# 7 7700X; 62 GB RAM. MoE expert weights pinned to CPU with an explicit
# --override-tensor regex (not --n-cpu-moe). Replaced the HauhauCS
# Qwen3.6-35B-A3B model on 2026-08-24.
# Qwen3.6-35B-A3B model on 2026-08-24. Switched from CRACK Q8_0 to Q6_K on
# 2026-08-25.
#
# WARNING: do not name any variable TEMP, HOST, or GPU_MAX_HEAP_SIZE. The
# ROCm HIP runtime reads those bare env vars (see the 2026-08-13 TEMP
# incident). This box is NVIDIA/CUDA, but the rule keeps the env portable.
# Model files (Q8_0 CRACK GGUF + f16 mmproj vision encoder + MTP head bundled
# Model files (Q6_K CRACK GGUF + f16 mmproj vision encoder + MTP head bundled
# in the GGUF). Dealign2026-08-24.
MODEL_DIR=/home/dank/models/Ornith-1.5-35B-A3B-CRACK
MODEL_FILE=Ornith-1.5-35B-A3B-CRACK-Q8_0.gguf
MODEL_FILE=Ornith-1.5-35B-A3B-CRACK-Q6_K.gguf
MMPROJ_FILE=mmproj-Ornith-1.5-35B-A3B-f16.gguf
# Upstream llama.cpp binary (b10121).
LLAMA_SERVER=/home/dank/llama.cpp/build/bin/llama-server
# Qwen-Sharp chat template (tracked in the repo, staged on the host).
# Qwen-Sharp chat template from peculiar-ragdoll (tracked in the repo,
# staged on the host).
# Source: https://huggingface.co/peculiar-ragdoll/Qwen-Sharp-Chat-Templates
CHAT_TEMPLATE_FILE=/home/dank/serverAppConfigs/scripts/chat-template.jinja
ALIAS=frank-default-small
BIND_ADDR=0.0.0.0
LISTEN_PORT=7770
# 3 parallel slots x 229376 ctx each (262144 lowered by 32k per stream for
# cache headroom) = 688128 total context pool. The per-slot 229376 matches
# the per-slot context advertised in models.json (229376).
CTX_SIZE=688128
# 3 parallel slots x 196608 ctx each (262144 lowered by 64k per stream for
# cache and checkpoint headroom) = 589824 total context pool. The per-slot
# 196608 matches the per-slot context advertised in models.json (196608).
CTX_SIZE=589824
PARALLEL=3
# Spec: MoE model with a bundled MTP head -> native draft-mtp, n_max 4
@@ -38,11 +41,12 @@ SPEC_DRAFT_N_MAX=4
SPEC_DRAFT_N_MIN=0
SPEC_DRAFT_P_MIN=0.0
# Expert offload via explicit tensor override (replaces --n-cpu-moe). Only
# the MoE FFN exps tensors move to CPU; everything else up to --gpu-layers
# stays on the 24 GB card. The regex is the same one the built-in --cpu-moe
# expands. Buffer type "CPU" is the CPU buft name on this host.
OVERRIDE_EXPERT='\.ffn_(up|down|gate|gate_up)_(ch|)exps'
# Expert offload via explicit tensor override (replaces --n-cpu-moe). The Q6_K
# layer audit found 630 MiB of expert tensors per block. Keep the four
# output-adjacent blocks 37-40 on CUDA, including the bundled MTP block 40.
# Move expert tensors from blocks 0-36 to CPU. All non-expert tensors stay on
# the 24 GB card. Buffer type "CPU" is the CPU buft name on this host.
OVERRIDE_EXPERT='blk\.(0|1|2|3|4|5|6|7|8|9|10|11|12|13|14|15|16|17|18|19|20|21|22|23|24|25|26|27|28|29|30|31|32|33|34|35|36)\.ffn_(up|down|gate|gate_up)_(ch|)exps'
OVERRIDE_BUFT=CPU
N_GPU_LAYERS=all
scripts/frank-default-small.service
+1 -1
View File
@@ -1,5 +1,5 @@
[Unit]
Description=llama.cpp Ornith-1.5-35B-A3B CRACK Q8_0 MTP + Sharp template (dankserver, frank-default-small)
Description=llama.cpp Ornith-1.5-35B-A3B CRACK Q6_K MTP + Qwen-Sharp template (dankserver, frank-default-small)
After=network-online.target
Wants=network-online.target
scripts/serve-frank-small.sh
+5 -7
View File
@@ -1,16 +1,14 @@
#!/usr/bin/env bash
# Serve dealignai/Ornith-1.5-35B-A3B-UNCENSORED (CRACK, Q8_0) as
# Serve dealignai/Ornith-1.5-35B-A3B-UNCENSORED (CRACK, Q6_K) as
# frank-default-small via upstream llama.cpp on Dankserver (RTX 3090, MoE
# experts on CPU).
#
# Runtime: upstream llama.cpp llama-server (b10121). The model is a 35B-total /
# 3B-active MoE-dense hybrid with the MTP speculative head bundled in the GGUF.
#
# Expert offload: instead of the layer-counter --n-cpu-moe, the expert (FFN
# exps) tensors are pinned to CPU explicitly with --override-tensor. The
# pattern is the internal MoE exps regex (the same one --n-cpu-moe expands),
# so only the Mixture-of-Experts weight tensors move to CPU; every other
# tensor stays on the GPU up to --gpu-layers.
# Expert offload: the expert (FFN exps) tensors selected by the layer-bounded
# --override-tensor pattern are pinned to CPU. The final output-adjacent
# blocks stay on CUDA with all non-expert tensors, including the MTP head.
#
# Spec: MoE model with a bundled MTP head -> native draft-mtp (the head is
# auto-detected from the target model; b10121 creates the MTP draft context
@@ -32,7 +30,7 @@ fi
# --- defaults (used when the env file does not set them) -----------------
: "${MODEL_DIR:=/home/dank/models}"
: "${MODEL_FILE:=Ornith-1.5-35B-A3B-CRACK-Q8_0.gguf}"
: "${MODEL_FILE:=Ornith-1.5-35B-A3B-CRACK-Q6_K.gguf}"
: "${MMPROJ_FILE:=mmproj-Ornith-1.5-35B-A3B-f16.gguf}"
: "${LLAMA_SERVER:=/home/dank/llama.cpp/build/bin/llama-server}"
: "${CHAT_TEMPLATE_FILE:=$SCRIPT_DIR/chat-template.jinja}"