i BSrSSKr\R"S5rSSS.SSS S S S .S SSSSS.SS01Sk1Sk/r0r0r/r\Hr\ "\\ 5(au\R\5 \R\R5VVs0sHupX_M snn5 \R\"\5\"\R55-5 M\R\"\55 M \"\5\"\5-\S-\S-r\"S\55r\"S\55rgs snnf)zVoice-design instruct constants for TTS inference. Defines speaker attribute tags (gender, age, pitch, accent, dialect) and translation/validation utilities between English and Chinese. Used by ``OmniVoice.generate()`` for voice design mode. Nz[\u4e00-\u9fff]u男u女)malefemaleu儿童u少年u青年u中年u老年)childteenagerz young adultz middle-agedelderlyu 极低音调u 低音调u 中音调u 高音调u 极高音调)zvery low pitchz low pitchzmoderate pitchz high pitchzvery high pitchwhisperu耳语> indian accent korean accentbritish accentchinese accentrussian accentamerican accentcanadian accentjapanese accentaustralian accentportuguese accent> 东北话 云南话 四川话 宁夏话 桂林话 河南话 济南话 甘肃话 贵州话 陕西话 青岛话 石家庄话c#^# UH#n[RU5(aMUv M% g7fN_ZH_REsearch.0is 8/mnt/workspace/OmniVoice/omnivoice/utils/voice_design.py r*As V*=QV]]STEUqq*=- -c#^# UH#n[RU5(dMUv M% g7fr"r#r&s r)r*r*BsR*=QqAQqq*=r+)__doc__recompiler$_INSTRUCT_CATEGORIES_INSTRUCT_EN_TO_ZH_INSTRUCT_ZH_TO_EN_INSTRUCT_MUTUALLY_EXCLUSIVE_cat isinstancedictupdateitemsappendsetvalues_INSTRUCT_ALL_VALID frozenset_INSTRUCT_VALID_EN_INSTRUCT_VALID_ZH)kvs00r)rBs^$ & ' e$HX3%K"+&(PV"! D$!!$'!!DJJL"ALDA14L"AB$++CIDKKM8J,JK$++CI6 !c"455"2&'"2&' V*=VVR*=RR#Bs>D