Commit Graph

3 Commits

Author SHA1 Message Date
vh 377f8a43c8 docs(reranker): record cutover VERIFIED + v13 gate PASS, A2 teardown, throughput
Brokkr independent verify clean (maxdiff 0.000000, no split). R42 v13
acceptance gate PASSES first time in its history: main+kb 56/90->90/90,
evictions 33->0. A2 control torn down. A3 throughput characterized at
~34 req/s (graceful queueing), with A4/util-bump/replica as levers.
2026-08-06 10:49:50 -07:00
vh 2c11748f87 chore(reranker): harden A3/A4 backends restart=unless-stopped (reboot survival)
A3 now backs the prod reranker alias but was launched --restart no;
docker update to unless-stopped so an ana-ml2 reboot can't silently
break the alias. Full compose-service promotion tracked as a follow-up
in the selection ledger.
2026-08-06 10:42:45 -07:00
vh ad2df89c0c feat(litellm): repoint fleet reranker alias to bge-reranker-v2-m3 (Brokkr R43)
The incumbent Qwen3-Reranker-0.6B was measured actively harming 80/90
fleet queries on main+knowledge_base (and inverting the bare-name region
behind Worldtree #389) — no-reranker beat it 89/90 vs 56/90. Brokkr's R43
bake-off selected BAAI/bge-reranker-v2-m3 (A3): 90/90 top-10, mean rank
0.19, multilingual (XLM-R), ~1.2 GB lighter than the incumbent.

Control arm (A2 = same Qwen weights, seq-cls head) scored identical to the
incumbent, proving the fault is a training prior, not the serving head —
which cancelled the expensive Qwen3-4B arm before it cost a GPU seat.

Cutover boundary 2026-08-06T17:37:48Z. The qwen3-reranker alias and the
:8002 backend are retained for one-edit rollback. Adds the process audit
trail at docs/pfi/reranker-selection-ledger.md.
2026-08-06 10:40:26 -07:00