frank-default-small: swap to batsclamp Huihui-Qwen3.6-35B-A3B-abliterated-FP8 on FreeToken

- same qwen3_5_moe arch and native block-fp8 format (e4m3, 128x128
  weight_block_size, un-fused per-expert layout) as the official
  Qwen FP8 checkpoint: identical FreeToken loader path, pool numbers
  unchanged (2048 MoE slots / 453005 KV tokens / 4 requests)
- checkpoint downloaded to /home/dank/models/huihui-qwen36-35b-a3b-abliterated-fp8
  (26 shards, ~35G); official checkpoint kept for rollback
- ships froggeric fixed Qwen3.6 chat template, read from checkpoint
- also refresh both .service Descriptions (large still said STRIX_LEAN)

scripts/frank-default-large.service

@@ -1,5 +1,5 @@
 [Unit]
-Description=ROCmFPX llama.cpp Qwen3.8-Flash-Next ROCmFP4 STRIX_LEAN 56B MoE (dankclaw, frank-default-large)
+Description=ROCmFPX (kingjones30 fork) llama.cpp Qwen3.8-Flash-Next ROCmFP4 STRIX 177B MoE (dankclaw, frank-default-large)
 After=network-online.target
 Wants=network-online.target
 

scripts/frank-default-small.env

@@ -1,5 +1,5 @@
-# Tunables for frank-default-small (Qwen3.6-35B-A3B-FP8,
+# Tunables for frank-default-small (batsclamp/Huihui-Qwen3.6-35B-A3B-
-# qwen3_5_moe hybrid MoE, Dankserver, RTX 3090).
+# abliterated-FP8, qwen3_5_moe hybrid MoE, Dankserver, RTX 3090).
 #
 # Runtime: FreeToken v0.1.2 (edge-native MoE serving engine).
 # Installed via uv into /home/dank/freetoken/.venv (CPython 3.12).
@@ -8,13 +8,18 @@
 # layers page KV (~20 KiB/token); the 30 GDN layers use constant-size
 # per-sequence state (mamba slot pool).
 #
-# Model: Qwen/Qwen3.6-35B-A3B-FP8 (HuggingFace safetensors, modelopt
+# Model: batsclamp/Huihui-Qwen3.6-35B-A3B-abliterated-FP8 (HuggingFace
-# block-fp8 quant, 128x128 weight_block_size, 35 GB on disk).
+# safetensors).  Abliterated (refusal-removed) Huihui finetune of
-# Architecture: qwen3_5_moe — 40 layers (10 attention + 30 GDN),
+# Qwen3.6-35B-A3B, quantized to the SAME native format the previous
-# 256 experts / 8 used, 2 KV heads × 256 dim, full_attention_interval=4.
+# official checkpoint used: modelopt block-fp8 e4m3, dynamic
-#
+# activations, 128x128 weight_block_size, bf16 weight_scale_inv;
-# Replaced the llama.cpp Tiel-Coder-35B-A3B (UD-Q6_K_XL) deployment on
+# lm_head/embeds/norms/router gates/visual/MTP kept BF16.  Un-fused
-# 2026-09-02.  The old serve script and env are in .AGENTTRASH/.
+# per-expert layout matches the official Qwen FP8, so the FreeToken
+# loader path is identical; ~35 GB on disk.  Ships the fixed Qwen3.6
+# chat template (froggeric/Qwen-Fixed-Chat-Templates) — FreeToken
+# reads it from the checkpoint; no template override needed.
+# Replaced Qwen/Qwen3.6-35B-A3B-FP8 on 2026-08-30; pool numbers
+# unchanged — same arch, same weight mass.
 #
 # Pool sizing: GPU budget (~17.6 GiB after weights + fixed caches) is
 # split MoE-first.  MOE_CACHE_SIZE pins resident expert slots; NUM_TOKENS
@@ -27,8 +32,8 @@
 # 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 directory (safetensors checkpoint, 35 GB).
+# Model directory (safetensors checkpoint, ~35 GB, block-fp8).
-MODEL_DIR=/home/dank/models/qwen36-35b-a3b-fp8
+MODEL_DIR=/home/dank/models/huihui-qwen36-35b-a3b-abliterated-fp8
 
 # FreeToken venv (uv-managed CPython 3.12, ft CLI, flashinfer, torch+cu130).
 FT_VENV=/home/dank/freetoken/.venv

scripts/frank-default-small.service

@@ -1,5 +1,5 @@
 [Unit]
-Description=FreeToken Qwen3.6-35B-A3B-FP8 qwen3_5_moe hybrid MoE (dankserver, frank-default-small)
+Description=FreeToken Huihui-Qwen3.6-35B-A3B-abliterated-FP8 qwen3_5_moe hybrid MoE (dankserver, frank-default-small)
 After=network-online.target
 Wants=network-online.target
 

scripts/serve-frank-small.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
-# Serve Qwen/Qwen3.6-35B-A3B-FP8 as frank-default-small via FreeToken
+# Serve batsclamp/Huihui-Qwen3.6-35B-A3B-abliterated-FP8 as
-# on Dankserver (RTX 3090, 60 GB RAM).
+# frank-default-small via FreeToken on Dankserver (RTX 3090, 60 GB RAM).
 #
 # Runtime: FreeToken v0.1.2 (edge-native MoE serving engine).
 #   - Experts live in host RAM; an LRU cache of expert slots is resident
@@ -10,10 +10,14 @@
 #     hold constant-size per-sequence state (mamba slots).
 #   - OpenAI-compatible API (/v1/chat/completions, /v1/models).
 #
-# Model: Qwen/Qwen3.6-35B-A3B-FP8 (safetensors, modelopt block-fp8,
+# Model: batsclamp/Huihui-Qwen3.6-35B-A3B-abliterated-FP8 — Huihui
-# 128x128 weight_block_size).  Architecture: qwen3_5_moe — 40 layers,
+# abliterated finetune of Qwen3.6-35B-A3B in the same native modelopt
-# full_attention_interval=4 (10 attention + 30 GDN), 256 experts / 8 used,
+# block-fp8 format (e4m3, 128x128 weight_block_size, un-fused per-expert
-# 2 KV heads × 256 dim.  35 GB on disk.
+# layout) as the official FP8 checkpoint this service ran before; the
+# FreeToken loader path is unchanged.  40 layers (10 attention + 30
+# GDN), 256 experts / 8 used, 2 KV heads × 256 dim.  ~35 GB on disk.
+# The checkpoint ships the fixed Qwen3.6 chat template; FreeToken reads
+# it from the model directory.
 #
 # Pool sizing: the GPU budget is split MoE-first between expert residency
 # and KV pages.  --moe-cache-size pins the number of resident expert
@@ -35,7 +39,7 @@ if [[ -f "$ENV_FILE" ]]; then
 fi
 
 # --- defaults (used when the env file does not set them) -----------------
-: "${MODEL_DIR:=/home/dank/models/qwen36-35b-a3b-fp8}"
+: "${MODEL_DIR:=/home/dank/models/huihui-qwen36-35b-a3b-abliterated-fp8}"
 : "${ALIAS:=frank-default-small}"
 : "${BIND_ADDR:=0.0.0.0}"
 : "${LISTEN_PORT:=7770}"