Run the routing heads in Go and route with them #206

2026-08-08T20:34:00+02:00

claude commented

2026-08-08 20:34:00 +02:00

Closes Vikunja #664.

Acceptance criteria and quality gate are in TASK.md on this branch.
Review the review, not the diff — leave comments and the agent will apply them via task start 664.

Closes Vikunja #664. Acceptance criteria and quality gate are in `TASK.md` on this branch. Review the review, not the diff — leave comments and the agent will apply them via `task start 664`.

claude added 6 commits 2026-08-08 20:34:00 +02:00

The tokenizer read every long word backwards (V-664) feabf9f350

encodeWord backtracks the Viterbi path from the end of the word and
prepends each piece, which puts them back in reading order. A second
reverse after that loop undid it. So "query: вода" tokenized to
[0 12 1294 41 12489 2] where the reference tokenizer gives
[0 41 1294 12 12489 2], and every multi-piece Russian word reached the
model with its pieces in the wrong order.

Measured on the recall fixture, same 27 cases either way:

  recall@1  70.4% -> 77.8%
  recall@3  85.2% -> 96.3%
  answered after gate  63.0% -> 66.7%
  false recall  0/5 -> 1/5

The classifier barely moves, 76.0% to 75.0% on the routing fixture,
because seeds and queries were mangled the same way and cosine survived
it. Recall is where it cost, because a stored passage and a live query
are different lengths and break differently.

The embedder id now names a tokenizer revision. Stored vectors were
written under rev 1 and no longer sit in the same space as a query
embedded now, and the model file's name never moved, so nothing would
have triggered ReembedAll.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013ptwopxyo3Z2kwFckHkLvN

Load the routing heads and read three of the four (V-664) 88c086482e

The heads trained in V-661 ran nowhere. This loads the exported graph
and reads intent, destination and clarify off one forward pass. It
declines below 0.6 max softmax rather than clarifying, so a declined
turn reaches whatever is behind it.

The slot head is exported and deliberately not read: slots already
come from the stage-2 extractor, and mapping BIO tags back to text
needs character offsets the tokenizer does not keep.

The clarify head decides on its own and decides first. It answers a
different question from the intent head, so a low intent confidence is
no reason to discard it. Reading it only above the intent threshold
cost 6 of the 8 ambiguous cases on the fixture: the word for water
reads as intent act at 0.23 and clarify at 0.98.

0.6 is the knee measured on the intent fixture: every higher value up
to 0.9 drops right answers and keeps the same two wrong ones.

The body is a fine-tuned COPY of the resident embedder and must never
replace it, because memory recall depends on that file scoring what it
scored.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013ptwopxyo3Z2kwFckHkLvN

Wire the heads between stage 0 and the resident model (V-664) 68a3c85186

They run before the model because they are two orders of magnitude
faster and score better on both halves of the route. They decline
rather than clarify, so a declined turn carries on to the model and
then the classifier, which is what a box with no weights file does on
every turn. Nil heads are byte-for-byte the cascade that shipped
before this.

Measured on the 96-case fixture, classifier+ONNX either way:

  intent       76.0% -> 96.9%
  destination  36.4% -> 75.8%
  false clarify   0 -> 1
  missed clarify  8 -> 1
  p50          24.5ms -> 27.9ms

That beats the gemma-4-12b cascade on both halves, 84.4% and 72.7%, at
a twelfth of its 329ms. The four remaining destination misses are all
calendar, which is the stage 0 trade V-660 flagged and the owner has
not called yet.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013ptwopxyo3Z2kwFckHkLvN

Give the daemon a heads_path and a fixture arm (V-664) a4abcdefa3

embedder.heads_path is empty by default and deploy/mavend.json sets
it. A missing or broken weights file logs and leaves the heads nil,
because refusing to start over a routing accelerator would trade a
working box for a better one.

TestONNXRoutingHeads is the same cascade TestONNXBaseline scores with
one arm added, so the two are directly comparable. It also checks the
Go tokenizer against the Python one, since the heads were trained
through transformers and are read through a hand-written tokenizer: a
mismatch shows up here as a score below what Python measured on the
same weights, and nowhere else. That is how the reversed word pieces
were found.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013ptwopxyo3Z2kwFckHkLvN

Record what the routing heads score in Go (V-664) 83e168f326

Two defects were found on the way: the tokenizer read every long word
backwards, and the clarify head was discarded below the intent threshold.
Both numbers are in the doc.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013ptwopxyo3Z2kwFckHkLvN

The routing heads run in Go now, so say so (V-664) 7138086c3f

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013ptwopxyo3Z2kwFckHkLvN

claude merged commit 7c77a378c1 into master

2026-08-08 23:21:41 +02:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: kami/Maven#206