Frank
a34322aeb0
The managed-skill library was regenerated: 94 narrow per-incident skills
replaced by 25 broader ones (radiance-*, llamacpp-*, homepage-*, fleet/
endpoint ops). Git renders this as mass delete + mass add because whole
directories changed name; old content stays in history at 12c5e88.
Surface check: the 25 on disk are exactly the skills the harness advertises
this session.
2.5 KiB
2.5 KiB
name, description
| name | description |
|---|---|
| radiance-w4-campaign | Run a radiance GPTQ-w4 quantization campaign on dankserver: calibration → protected experts → convert → KLD gate → deploy, with the engine-pin and MNBT laws. |
Pipeline for quantizing a BF16 Qwen3.8-Flash-Next checkpoint into a servable qwen4exp .rad on dankserver (frank-primary). Reference run: serverAppConfigs-docs/LLM/frank-primary-radiance-swift15-w4dn8.md; scripts in scripts/ (run-swift15-bf16.sh, calib-push.py, derive-protected-experts.py, apply-protected-cast.sh, convert-swift15-w4dn8.sh, kld-gate-swift15.sh, fleet-probes-swift15.sh).
Sequence:
- Fetch checkpoint with per-shard LFS sha256 verification (
-f, byte counts, no token on argv). - BF16 container:
rad-convertno-recipe (pure copy) — serves as the KLD reference engine. - KLD reference: record logits on the SERVE pin (1.0.13) at MNBT 32768,
--max-model-len 16384, corpus ≤16k prompts. Writes atomically at completion (rows:line = receipt). Never scp a running bash script (lazy-read law). - Calibration: engine 1.2.0 (only pin with
RADIANCE_CALIB_TOKENS); push ~10M tokens of the model's own distribution; per-expert down-proj Grams land as RADGRAM2 files (gram.r<rank>.blk.L.ffn_down_exps.<e>.bin). Grams are engine-invariant across pins — the version boundary is by design. - Protected experts: derive from gram energy (share ≥3 % AND ≥3× routing share), splice between
# PROTECTED-BEGIN/ENDmarkers of the recipe. Plugin law: ONE encoding per expert table; only whole bf16 experts legal beside it — no gate_up/down split. - Final convert on the SERVE pin (CPU GPTQ, hours; disk gate ≥146 GiB free first).
- KLD gate on the serve pin: MNBT ≤ 2048 (w4 streams in place only there; 32768 →
DID NOT FIT ... stored form is not the container's), pool 40960, fp8 KV matching the production KV dtype. Gate: mean KL ≤ 0.03, top-1 ≥ 95 %. - Deploy
./scripts/deploy.sh dankserver --service frank-primary; verify/v1/modelsid = served-model-name (alias independent of file — harness models.yml untouched), boot ledger, streamed ≥128-token decode probe, fleet probes (vision needs ≥256-token budget +reasoning_contentfallback).
Laws: stop frank-primary for every GPU phase (relinked unit: stop, mask fails); verify docs against podman inspect <name> --format {{join .Config.Cmd " "}} (rootless, as dank) before writing shapes; ff-pull fails on untracked files the merge tracks — trash first.