memory: run 6 TRAINING on pfi-gx10 (abliterated base 33c59654), trial alias dark, gate naming

This commit is contained in:
2026-09-08 04:26:14 -07:00
parent 3fec668bf2
commit 304baddfc1
+15 -3
View File
@@ -1,6 +1,6 @@
# Persistent memory — eshpfi-management
_Last updated: 2026-09-08 (fleet-ops: ERP run-5 COMPLETE + gated RESCUED on pfi-gx10 [landmark, first capability-gate pass in the line], WhereTF card+DNS, ext-tts alias fix, irv-ml1 stale-IP cleanup + ana-ml2 discovery proxy, Miranda relay authority)_
_Last updated: 2026-09-08 11:30Z (fleet-ops: **ERP run 6 TRAINING on pfi-gx10** on the first genuinely abliterated base [jenerallee78 ARA @ 0631379a, index 33c59654], run-5 seat unloaded; earlier today: run-5 RESCUED, WhereTF card+DNS, ext-tts alias fix, irv-ml1 stale-IP cleanup + ana-ml2 discovery proxy, Miranda relay authority)_
> **Always check for `/tmp/infra-ops-handoff.md`** — if it exists and its
> `Written:` stamp is under an hour old, read it (it carries the in-flight
@@ -144,8 +144,19 @@ COMPLETE and gated RESCUED (02:13 PDT). Live open items:_
pulled was stock google, and the `-heretic` name rode along from intent. **Operator/Brokkr decision
now evidenced (not a label guess): accept RESCUED-on-stock, or swap to a real abliteration (needs a
bf16 source, not the GGUF) + re-run. Recipes should drop `-heretic` from the base name.**
- **⏳ OPEN (operator + brokkr, not urgent): accept RESCUED-on-stock as the ERP seat, or source a bf16
abliteration + re-run the line.** Now an evidenced choice (base is stock). No deadline.
- **✅ RESOLVED → 🔥 RUN 6 TRAINING on pfi-gx10 (launched 2026-09-08T11:17:43Z, pid 4100375, `~/erp-tune/run-06.pid`,
524 steps @ ~4050 s/it → adapter ~18:3019:00Z).** Operator picked the abliteration: jenerallee78's ARA
abliteration of Gemma-4-26B-A4B-it @ `0631379a`, pulled to
`/home/infra-ops/models/gemma4-26b-a4b-it-ara-abliterated-jenerallee78-0631379a` (named for the bytes), 32/32
shards sha256+size == brokkr's pins, **base identity = index sha `33c59654…`**. Run-5 recipe byte-held (recipe-r6 =
r5 targets, survivors-r5 verbatim); the free check passed EXACTLY (encode 8,370 rec / ctx 18,598,779 / loss
9,935,076). Grant `operator-2026-09-08-rnd-run6` = the operator's direct in-session words ("unload the gx10 and
commence training on the gx10. window is open now."). ⚠ The repo's `tokenizer.json` bakes in a 256-token
truncation — STOCK tokenizer set installed over it, repo originals kept as `*.repo`. ⚠ `hf download` silently
ignores `--include` when given multiple patterns after one flag (one `--include` per pattern). **erp-tune-v5
seat STOPPED (GPU cleared for training) → LiteLLM `trial` alias is DARK until the next serve.** Gate choreography
when the adapter lands (brokkr, cc channel, hands-off): serve base as **`erp-seat-base-ara`** (NOT
`erp-seat-base` = stock), then merged arm as **`erp-tune-v6`**, run-5 flags. → `docs/runbooks/gx10-run-06.md`.
- **📮 althing reachability on THIS bg seat = the cc-channel route, NOT the waiter.** `althing-listen`
reaps within ~seconds-to-minutes on an idle background seat (harness kills detached bg tasks; hit it
repeatedly this session). Fix (operator-directed 2026-09-08): `althing-route declare --handle infra-ops
@@ -575,6 +586,7 @@ below is a live commitment or a known-open risk._
- `[2026-09-08]` **ERP run 5 = RESCUED (landmark R49.5)** — first capability-gate pass in the ERP-seat line; the 3.46%-loss dependency-forcing slot (GovReport+QMSum) broke the coupling runs 3c/4 couldn't. Seat `erp-tune-v5` served on gx10:8098, `trial` alias repointed 3c→v5. → `persistent-memory.d/2026-09-08-run5-rescued.md`
- `[2026-09-08]` **R47 base settled from bytes = STOCK `google/gemma-4-26B-A4B-it`** — three-way sha match (local == HF etag == stock LFS oid; commit `4d7ae498` == stock HEAD); the `-heretic` label is a naming error, all runs trained from stock. Accept-vs-swap now evidenced. → `persistent-memory.d/2026-09-08-base-provenance-stock.md`
- `[2026-09-08]` **ERP run 6 LAUNCHED on pfi-gx10 on the jenerallee78 ARA-abliterated base** (index `33c59654…`, 32/32 shards byte-verified vs brokkr pins, stock tokenizer set installed over the repo's 256-token-truncating one, run-5 recipe byte-held, free check exact). Operator's direct grant `operator-2026-09-08-rnd-run6`; run-5 seat unloaded (`trial` dark). Gate names: `erp-seat-base-ara` / `erp-tune-v6`. → `docs/runbooks/gx10-run-06.md`, commit `3fec668`.
- `[2026-09-08]` **Miranda = operator's chief of staff, may relay his directives** — added to user-level `~/.claude/CLAUDE.md` (dotfiles `7134a22`) as the named exception to the no-relayed-auth rule (unidentified peer relays still excluded); material-consequence calls she relays stay the operator's own.
- `[2026-09-08]` **Fleet fixes shipped** — WhereTF Homepage card + DNS (`4506ef6`); ext-tts LiteLLM alias → `irv-ml1.nh3.internal` (DB `/model/update` + `extra_hosts`, `957c8f1`); the 09-06 irv-ml1 stale-IP trail repointed across 25 composes + services.yaml + ssh-target → DNS name (`e0d1c44`); Homepage `/api/services` outage fixed — ana-ml2 discovery via a socat proxy on ana-docker (`stacks/ana-ml2-proxy`, `913d2d2`, reversible).
- `[2026-09-07]` **Fleet internal TLS pattern shipped** — caddy (cloudflare-plugin build, `~/.local/bin/caddy-cf`, `fleet-tls-caddy.service`) on nh3-dev is the wildcard cert authority: publicly-trusted LE `*.nh3.phasefinal.com` via Cloudflare DNS-01, resolved internally by an AdGuard split-horizon rewrite (all 3 resolvers). `talk` self-terminates on :8092 with the trusted cert (operator's in-container-TLS ruling), renewal auto-synced by `fleet-tls-cert-check.timer`. Interstitial gone; secure-context+AudioWorklet verified via headless Chromium. Pattern + foot-guns (restart-disrupts-inflight → clients need retry; wildcard = name-only, never IP) → auto-memory `reference_fleet_internal_tls_pattern`.