Skip to content

docs(WO67): controlled matrix — the schedule, not the corpora #349

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ronaldtse merged 1 commit into from
Oct 22, 2026
Merged

docs(WO67): controlled matrix — the schedule, not the corpora #349

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 9 additions & 0 deletions TODO.sota/67-b6-thai-lexicon.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,15 @@ by an old engine defect, and thai_words/volubilis were never tapped).
Fix in flight: tha-phonemic-v2 (122,311 pairs, corpora/tha-phonemic-v2.yaml,
models#343) training as run-080-thai-lex-v2; wiki-vocabulary
self-expansion (two-engine cross-check) building as run-080-thai-wiki-4.

**Controlled-eval matrix (2026-10-23, v1 test split = 3,225 ORST rows,
PER over phoneme tokens):** v1 22.56 | v2 (110k pairs, 3ep BS32)
**39.1** | v3 (131k incl. wiki tier, same schedule) **39.5**. More
data at a 5x-smaller update budget is WORSE like-for-like; both
big-data arms degrade identically -> the schedule is the confound,
not the corpora. run-080-thai-lex-v4 (v3 data, BS4 x 2ep ~= 65k
updates, matching v1's budget) is the control; wiki tier's own value
shows only after the schedule is fixed.
RTGS encodings landed maps#204; IPA layers = WO82.
Lexicon phase DONE: corpora/tha-phonemic.yaml -- 28,644 train / 3,225
test (word -> phonemic-with-tones pairs, ORST vocabulary x pythainlp
Expand Down
Loading