dank/serverAppConfigs
1
0
Fork 0
Code Issues 6 Pull Requests Actions Packages Projects Releases Wiki Activity
1,280 Commits 42 Branches 8 Tags
4d41366f4c5cf4c0124b46fe401f9ba4119c9832
Commit Graph
9 Commits
Author SHA1 Message Date
Frank 55f87fcdce llm: retire the trio service names — fleet services renamed to frank-archon/frank-dragoon/frank-zealot 
The 2026-10-09 wire aliases become the service names everywhere:
systemd units, serve scripts, env files and FRANK_* variables,
deploy-plan/deploy-remote service keys, container names, Prometheus
job names + alert rules, Grafana dashboard queries, bats seams,
runbook filenames, homepage/AGENTS lore, and historical prose.
Rollback tags renamed in lockstep. Residue in the tree: zero.
Retired frank-medium/small/npu names are history and stay.

Known casualty: scripts/patches/*.patch context lines were rewritten;
those retired-era vLLM-fork patches no longer apply byte-exact.
 2026-10-10 17:03:14 -07:00
Frank 1aef7418ad swift15 cutover: AGENTS.md + runbook + log + fleet probes + KLD gate shape fix 
- AGENTS.md: fleet table row + frank-primary section rewritten for the
  GPTQ w4 chain (1.0.13 serve/convert, 1.2.0 calib, int8 KV, vram pair).
- runbook: throughput gate PASS (136.0 t/s median @ 44.5k, 4.5x bar),
  fleet probes PASS, control baselines recorded.
- log.md: 2026-10-08 campaign entry with the four measured engine laws.
- kld-gate: shape matched to the reference recording, then reshaped for
  the w4 MNBT<=2048 file-tier law (pool 40960).
- fleet-probes-swift15.sh: tools/reasoning/response_format/vision walk.
 2026-10-08 04:41:07 -07:00
Frank 8375567de3 swift15: restore expert_tiered + disk-tier flags dropped in the gate rewrite 2026-10-08 03:18:04 -07:00
Frank 79c2e64c2c swift15: gate MNBT 3072 per Chief (w4 arena scales hard with MNBT; 32k step is absurd here) 2026-10-08 03:17:33 -07:00
Frank f20da6a6ff swift15: gate shape MNBT 16384 (w4 arena at 32768 costs 33 GiB), kv 2048, headroom 512 2026-10-08 03:16:45 -07:00
Frank 058ea6cca8 swift15: gate boots at the production pool (w4 gate_up experts are RAM-resident; disk tier refuses non-serving stored forms) 2026-10-08 03:14:20 -07:00
Frank 3f8531179e swift15: gate flags per engine (--kld-ref takes no corpus; --kld-out JSON parsed) 2026-10-08 03:12:18 -07:00
Frank 5191c49967 swift15: KLD gate script (candidate vs bf16 ref, KL<=0.03 / top-1>=95%) 2026-10-07 23:43:06 -07:00
Frank fa858f4c10 feat(campaign): protected-cast applier, recipe markers, KLD gate runner 
apply-protected-cast.sh splices derive-protected-experts.py output into
the recipe's PROTECTED-BEGIN/END section (refuses to re-derive, trashes
the pre-edit recipe). kld-gate-swift15.sh boots the candidate on the
serving pin in --kld-ref mode (pool fully resident, no disk tier) and
enforces mean KL<=0.03 / top-1>=95% from the JSON report.
 2026-10-06 22:41:18 -07:00