noonghunnaandClaude Opus 4.7 a2fc05e1c8 v0.8.0: doc generated composes are not relocatable (run with --project-directory) ...

Generated composes inherit the shipped service definition's repo-relative
bind-mount sources verbatim (synthesize-nothing model, §4). Document that they
must run with --project-directory anchored to the shipped compose dir; explain
the dangling-mount failure mode (Docker auto-creates empty dirs -> vLLM
"Is a directory" boot crash). Surfaced by the on-rig golden-parity leg.

Co-Authored-By: Claude Opus 4.7 (1M context) <[email protected]>

2026-05-16 17:19:22 +00:00

..

assets/tq3-mtp-genesis

docs(tq3-mtp): add missing 04-gemma-vs-qwen.png chart

2026-05-16 04:54:05 +00:00

engines

docs: use \$MODEL_DIR placeholder, not the dev rig's /mnt/models/huggingface/

2026-05-10 17:13:29 +00:00

img

docs: refresh 4090 cross-rig knee with @laurimyllari's richer 38-cap sweep

2026-05-09 19:54:49 +00:00

ADDING_MODELS.md

docs: expand KV_MATH + add ADDING_MODELS workflow

2026-05-14 20:10:22 +00:00

ARCHITECTURE.md

Split charts per GPU-count page; chart sources land in tools/charts/

2026-04-29 14:42:09 +00:00

CI_RUNNER_SETUP.md

fix(ci): simplify vllm image workflow, drop smoke-gate (#135)

2026-05-14 23:03:02 +00:00

CLIFFS.md

docs(soak-test): clarify PASS verdict semantics — closes #140

2026-05-15 22:18:09 +05:00

COMPARISONS.md

Restructure docs around hardware axis: SINGLE_CARD.md + DUAL_CARD.md

2026-04-29 13:44:16 +00:00

COMPOSE_GENERATOR.md

v0.8.0: doc generated composes are not relocatable (run with --project-directory)

2026-05-16 17:19:22 +00:00

CONTAINER_RUNTIMES.md

docs(runtimes): tighten Proxmox section — native venv works (#49)

2026-05-04 21:57:34 +00:00

DTYPE_MATRIX.md

docs(dtype-matrix): more polish — RDNA naming, FP8 maturity caveats, AMD detection

2026-05-12 12:16:40 +00:00

DUAL_CARD.md

feat: unify dual-card composes with NVLink auto-detection

2026-05-14 02:36:15 -02:30

EXAMPLES.md

feat(bounded-thinking): Phase 3 grammar A/B complete; DeepSeek scratchpad is the new recommended grammar

2026-05-04 11:18:08 +00:00

FAQ.md

docs: update launch.sh references for v0.6.1 wizard flow

2026-05-14 00:00:39 +00:00

GLOSSARY.md

docs(dtype-matrix): per-arch hardware accelerator matrix for compose optimization

2026-05-12 12:04:16 +00:00

HARDWARE.md

docs(HARDWARE): add note on PCIe Gen 3 + older CPU TP=2 headwind

2026-05-15 22:18:09 +05:00

INFERENCE_ENGINES.md

restructure: promote topology to a directory level (single/dual/multi4)

2026-05-09 12:11:17 +00:00

KERNEL_MATRIX.md

docs(KERNEL_MATRIX): add Kernel Selection Philosophy section

2026-05-15 23:11:50 +00:00

KV_MATH.md

docs(kv-math): extend k_v_tensors=N notation to sliding-KV formulas

2026-05-15 11:01:55 +00:00

MULTI_CARD.md

feat(launch): add hardware topology advisor

2026-05-15 09:54:04 +00:00

NIGHTLY_BUMP_RUNBOOK.md

reorg: services/ consolidation + gpu-mode under git + ComfyUI + pin tracker + path updates

2026-05-10 16:57:03 +00:00

PATCH_ATTRIBUTION.md

v0.8.0 STEP 2: extract patch_attribution.py (sound body-only reaches(), test imports it)

2026-05-16 15:29:10 +00:00

PATCH_POLICY.md

v0.8.0 STEP 5: COMPOSE_GENERATOR.md + PATCH_POLICY.md (#141 contributor contract)

2026-05-16 15:57:02 +00:00

QUALITY_TEST.md

feat(gemma-4-31b): INT8 PTH KV unblocks 262K + AWQ + DFlash compose family

2026-05-10 16:56:17 +00:00

SINGLE_CARD.md

docs: update launch.sh references for v0.6.1 wizard flow

2026-05-14 00:00:39 +00:00

STRUCTURED_COT.md

restructure: promote topology to a directory level (single/dual/multi4)

2026-05-09 12:11:17 +00:00

TQ3_MTP_GENESIS.md

docs(tq3-mtp): writeup + charts for the Genesis-backed TQ3+MTP path

2026-05-11 20:32:43 +00:00

UPSTREAM.md

docs: UPSTREAM Gemma4 TurboQuant row — exact config.py:101 mechanism + fix-PR set

2026-05-16 04:49:37 +00:00