docs(wiki): close the session's loose threads — bay printer question, vocabulary-in-code decision, training note, guard family, OQ #16 settled

Claude-Session: https://claude.ai/code/session_01FWncR69HgGPuei1dLrW3cU
2026-09-06 22:37:57 +02:00
parent e67f0ccef0
commit ec44547122
5 changed files with 33 additions and 5 deletions
@@ -282,3 +282,11 @@ read. Composed `model_version` reads `<plate>+yolox:yolox_s.onnx@640`.
  maps to Vetura and no downgrade fires; vans, trucks, buses and motorcycles do separate. Phase B
  (a body-type classifier on the pilot's own frames — every wash order is a labelled frame) is
  what closes that gap; the detector's box is the crop it will classify.
+- **Training (user asked, 2026-09-06): YOLOX-S needs none.** It ships trained on COCO and is a
+  finished detector; its limit is vocabulary, not quality. Phase B is a *different, smaller*
+  model — a classifier over the detector's crop, not a retrained detector — fine-tuned on a small
+  Apache-licensed backbone. Data: a few hundred crops per category to start, a couple of
+  thousand is comfortable, all from the pilot's own lane and camera. Labels are NOT the
+  operator's picks (untrusted — [[threat-model]]) but a trusted reviewer's, gathered through the
+  [[vision-review-outbox]]. Expect 85–95 % on frontal gate views once tuned — enough to flag,
+  never to bill, which is why the flag records and the site threshold exists.