Skip to content

固定ratioモデルのincremental差分を分離して観測 - #42

Merged
PyYoshi merged 1 commit into
devfrom
v3/ratio-chunk-observations
Sep 20, 2026
Merged

PyYoshi merged 1 commit into
devfrom
v3/ratio-chunk-observations

Conversation

@PyYoshi

@PyYoshi PyYoshi commented Sep 20, 2026

Copy link
Copy Markdown
Owner

概要

  • 固定済み因子1/0.90とlegacyをwhole/1/7/64/1024-byteで比較。入力は4 KiB以下。
  • chunk差分、legacy同一feedとの差分、同profile因子1との差分を分離。
  • 候補数、順序、confidence bits、doneを保持。feed回数の必然的な差は不一致指標から分離。

観測

cp1252 decode-equivalentは1-byteで全model 0/8、7-byteでlegacy 6/8、生成model 4/8。wholeでは因子0.90が8/8でもincremental gateは達成していません。全結果と制約を日本語文書へ記録。係数やgridは変更しません。

検証

関連102 unittest成功、Ruff/diff check成功。全400観測を2回実行しreport全byte一致。標準targetのbuild間一致、wholeと固定sweep一致を要求。

標準model・C++・公開API変更なし。追加fuzz/大入力/P01は未実施。

関連: PyYoshi/cChardet#123

@PyYoshi
PyYoshi marked this pull request as ready for review September 20, 2026 23:53
@PyYoshi
PyYoshi merged commit 4138e44 into dev Sep 20, 2026
11 checks passed
@PyYoshi
PyYoshi deleted the v3/ratio-chunk-observations branch September 20, 2026 23:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant