feat(C-3): 旁白/章節切分模型遷 gemini-3.5-flash（A/B 驗過後切換）#80
Merged
Conversation
劉老師本機 A/B（PID 控制器三頁）驗過品質後拍板遷移：正確性兩邊乾淨、3.5 口吻更自然且長度更貼 75s/頁、成本約 64%；唯一弱點是偶爾漏列點，已補 prompt 糾正。 - slide_ingest.py：寫死 MODEL=gemini-2.5-flash → narration_model()=resolve_id(TEXT_FAST) （呼叫時解析，設定頁 text.fast 覆寫即時生效）。旁白 + 章節切分一起遷（共用模型常數）。 - NARRATION_PROMPT_DETAILED 補第 7 條：條列項目每項至少帶一句、可精簡不可整項遺漏。 - rollback 免改 code：設定頁把 text.fast 覆寫回 2.5 即可。 - solve.py（解題）不在此遷移範圍（正確性更敏感、未 A/B）。 本機全套 2702 passed（剩 1 QR 像素為容器缺 Noto CJK 字型假象，CI 權威）。
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
做什麼
劉老師本機跑 C-3 A/B（PR #79 的
tools/ab_narration.py，材力/自控Chap08-PID控制器設計三頁）驗過品質後拍板遷移。把slide_ingest.py的旁白 + 章節切分從寫死gemini-2.5-flash（將淘汰）切到 M 軸角色登錄表text.fast（=gemini-3.5-flash）。A/B 結論（為什麼切）
_truncate_at_sentence截 ✅怎麼改
slide_ingest.py：寫死MODEL = "gemini-2.5-flash"→narration_model() = resolve_id(TEXT_FAST)，呼叫時解析（設定頁text.fast覆寫即時生效）。旁白 + 章節切分一起遷（共用同一模型常數，章節切分也得離開 2.5）。計帳如實落 resolved id。NARRATION_PROMPT_DETAILED補第 7 條：「條列項目每項至少帶一句、可精簡不可整項遺漏」——補上 3.5 唯一弱點（漏列點）。docs/PRODUCT_READINESS.mdC-3 →[x]。範圍與 rollback
solve.py（解題）模型不在此遷移範圍——正確性更敏感、未 A/B，另議。text.fast覆寫回gemini-2.5-flash即時退（M-3model_roles最高優先）。怎麼測
narration_model()解析 =gemini-3.5-flash（驗證過）。本機全套 2702 passed（剩 1 QR 像素為容器缺 Noto CJK 字型假象，CI 權威）。不碰 review gate / 狀態機 / render 入口 assert——只換「旁白用哪個模型生」（硬規則 #1）。Reviewer 決策點
無——A/B 已由劉老師驗過、範圍與 rollback 如上。後續若要連
solve.py解題模型也遷，需另跑 A/B（正確性敏感）。Generated by Claude Code