Was '**v0.8.2 surfaces (current):**' — stale at v0.8.7. The four pull-gate
bullets are still the current surface, so reword to '**Pull-gate surfaces
(current):**' (no pinned version → can't go stale on the next release).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
ARCHITECTURE.md described the v0.8.0 pull→gate→emit→boot→loop pipeline but
not v0.8.2's user surfaces. Adds: the --recommend verdict UX, the
--submit-last/--submit failure on-ramp, the broadened arch registry
(native built-ins reach a clean serve verdict, per-repo remote-code still
fail-closed), and the optional non-NVIDIA hwdetect subprocess; scripts-tree
pull.sh line annotated. README universal-pull bullet refreshed
(v0.8.0; extended in v0.8.2 — recommend + on-ramp + wider coverage).
Docs only; no behaviour change.
User feedback: the runnable command sat ~40 lines below prose. Adds a
copy-paste Quickstart (the one command + the 3 outcomes/exit codes) at the
TOP of docs/PULL.md, and a one-liner in ARCHITECTURE.md clarifying the
[C0]/[B]/[C1] stage taxonomy is internal — users just run scripts/pull.sh.
Verified: the referenced `--list` flag does NOT exist (dropped — cited a
real COMPOSE_REGISTRY key + the Usage section instead); leak-clean; docs-only.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
ARCHITECTURE.md is the canonical current-state orientation doc but had no
mention of pull.sh / generate-compose.sh / the gate→emit→loop pipeline /
.pull-captures. Adds a "Two ways in: curated catalog + universal pull"
section (curated = measured backbone; pull = honest, never-silently-passes
front door) with the pipeline diagram + the boot-fit≠runtime caveat, and
updates the scripts/ tree. Present-tense current-state voice; leak-clean;
all links resolve. Ships with v0.8.0 under #147.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
SINGLE_CARD.md and DUAL_CARD.md embed scoped charts now — single-card
configs only on the single page, dual on the dual page. Combined views
stay on top-level README and the model README.
- docs/img/performance-{single,dual}.{svg,png} — new, scoped TPS charts
- docs/img/vram-budget-{single,combined}.{svg,png} — new
- docs/img/vram-budget-dual.{svg,png} — content swap: was combined,
now genuinely dual-only. Old combined content lives in -combined.
- tools/charts/gen-{perf,vram}.py — matplotlib sources, idempotent.
Re-run with: uv run --with matplotlib --with numpy python3 ...
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
User feedback: navigating to relevant docs was cumbersome. The natural
first decision is "1 GPU or 2 GPUs?" and the existing docs mixed model-
specific reference with deployment guidance.
New navigation:
- README.md adds a "Pick your path" pivot pointing at hardware-axis pages
- docs/SINGLE_CARD.md — 1× 3090 deployment menu (workload → compose →
TPS) with all single-card configs (vLLM + llama.cpp), VRAM budget,
prefill cliffs explained operationally, what single-card can't do
- docs/DUAL_CARD.md — 2× 3090 mirror (4 dual variants + TP=2 explainer +
what dual unlocks vs single + Marlin pad fork dependency)
Slimming:
- models/qwen3.6-27b/README.md: dropped duplicated variant tables
(now in GPU-count pages); kept model-specific content (quants, Genesis
patch surface table, what's working / not, VRAM diagram)
- models/qwen3.6-27b/USE_CASES.md: deleted. Per-workload content
absorbed into the GPU-count pages (deduplicated). Troubleshooting
list moved to docs/FAQ.md as a new "Troubleshooting" subsection.
Image-token cost / vision specifics absorbed into SINGLE_CARD.md.
Reference updates: 8 files updated (engines/VLLM.md, engines/README.md,
COMPARISONS.md, ARCHITECTURE.md, EXAMPLES.md, FAQ.md, INTERNALS.md,
top-level README) — all USE_CASES.md links re-pointed to SINGLE_CARD/
DUAL_CARD where appropriate.
Net delta: -167 lines (was 197 in USE_CASES + duplicated tables in
model README; now 342 lines split between SINGLE_CARD + DUAL_CARD with
content deduplicated against each other).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
