訂正Grammar models
Neural grammar correction, measured honestly
One model per language family — single-file ONNX, loaded through the gem, no Python runtime. Every score below was decoded from the exact artifact you download and scored by a pipeline built to reject inflated numbers.
English English
0.535
422 ms/sent · 4 rnd · RSS 704 MB
French Français
0.9803
Spanish Español
0.668
German Deutsch
Measured on its held-out suite with an LLM-judge gate; see the model card.
Russian Русский
0.8333
Chinese · Japanese · Korean 中文 · 日本語 · 한국어
0.661 / 0.788
Korean 한국어
0.394
Urdu اردو
0.9137
446 ms/sent · 1 rnd · RSS 1061 MB
Portuguese Português
0.571
Dutch Nederlands
0.565
Polish Polski
0.418
Italian Italiano
0.374
Ukrainian Українська
0.676
Turkish Türkçe
0.597
Serbian Српски
0.509
Greek Ελληνικά
0.367
| Language | Grammar correction | Spell-check tier | Grammar rules |
|---|---|---|---|
| English | yes | yes | yes |
| French | yes | planned | — |
| Spanish | yes | planned | — |
| German | yes | planned | — |
| Russian | yes | planned | — |
| Chinese · Japanese · Korean | yes | planned | — |
| Korean | yes | planned | — |
| Urdu | yes | planned | yes |
| Portuguese | yes | planned | — |
| Dutch | yes | planned | — |
| Polish | yes | planned | — |
| Italian | yes | planned | — |
| Ukrainian | yes | planned | — |
| Turkish | yes | planned | — |
| Serbian | yes | planned | — |
| Greek | yes | planned | — |
F0.5 weights precision 5× over recall — a proofreading tool must not invent errors. Protocols: CoNLL-14 via the official M² scorer; fr/es via the strict span scorer on held-out corruption benchmarks with artifact guards.
評価中In final evaluation
Five more languages at the gate
These models are trained and packaged; their scores are being measured now. Numbers appear when they survive the pipeline — not before.
測定The measurement law
Measured, not marketed
Every score on this page comes from the same frozen pipeline: decodes run against the exact artifact you download, and the scorer rejects the failure modes that inflate GEC numbers — output inflation, CJK tokenization mirages, span-size mirages. Where a comparison appears, the baseline was run on the same test set. Numbers that cannot survive this pipeline are not published.
Verified on Apple Silicon (arm64) and x86-64 Linux
Single-file ONNX per model, loaded through the gem with no Python runtime. Every shipped file passes a cross-architecture confidence check before publish.
gem install kotoshu
# fetch a grammar model into the gem cache
python scripts/fetch_grammar_model.py --lang fr # → ~/.cache/kotoshu/grammar/fr