monster/modelbeast
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

benchmarks: Qwen-Layered complete ladder — q8 floor (q6/q4 quality cliff), 1024 hero tier, operator shipped; flux_local audit (FLUX.1 compile gap)

Browse Source 
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
m3ultra 2026-07-19 22:20:49 +10:00
parent 58a2a8177d
commit d0b02003b8
1 changed files with 25 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
BENCHMARKS.md
View File

@ -373,3 +373,28 @@ MODELBEAST operator wrapper (qwen_layered_local) once defaults settle.
Both Ultras now serve Qwen-Image-Layered from the same 34GB baked-q8
artifact (m3: ~/qwen-layered/, m1: ~/qwen-layered-staging/) + the same
Gitea fork clone. Peak 36.7GB either box. m3 disk after install: ~82GB free.
### Qwen-Image-Layered — COMPLETE ladder + verdicts (2026-07-19/20)
| config | box | wall | verdict |
|---|---|---|---|
| q8 · 20 steps · 640 | m3ultra | **243.0 s** | **volume default** |
| q8 · 20 steps · 640 | m1ultra | 365.6 s | second lane |
| q8 · 20 steps · **1024** | m3ultra | 857.6 s | **hero tier** — superb detail, true RGBA (68% transp) |
| q6 · 20 steps · 640 | m1ultra | 366.4 s | ❌ REJECTED — decomposition semantics break (silhouette, no cutout) |
| q4 · 20 steps · 640 | m1ultra | 350.1 s | ❌ REJECTED — pure noise |
Quant findings: speed is FLAT across q4/q6/q8 (compute-bound at these
settings) — quantizing below q8 buys only disk/RAM and DESTROYS quality.
**q8 is the floor for this model.** q6/q4 bakes deleted (m1 back to 3.4T).
Operator `qwen_layered_local` shipped: vendored fork at
vendor/mflux-qwen-layered (install_qwen_layered.sh), tuned defaults
(20/640/baked-q8, resolution 1024 opt-in), m1 registered in nodes.json
(machine-local), e2e run.py contract smoke passed on m3.
Also audited flux_local (mflux 0.18.0): DEFAULT PATH HEALTHY — flux2-klein
family has mx.compile on denoise + compiled scheduler (the 9s/image path).
GAP: FLUX.1 family (dev/schnell/krea) has ZERO mx.compile — backport =
same env-gated pattern as Hunyuan T1/T2, modest expected win (1-4% M3U,
more on small boxes), upstreamable to mflux. Queued, not urgent.