{"text": "請評估本項目加入中英文支持。上期視頻我介紹了Audio CPP,當時我把它比作音頻版的Nama CPP,僅僅過了幾天,它又一次大升級,不僅一口氣新增了10個模型,還支持加載GGUF量化版,並且落地了牛市生成。甚至現在,你看到的這個語音輸入法,就是用它做的。操作介面上,基礎用法沒變,主要改動是增加了多語言支持,GGUF量化轉換,還有音色庫管理。接下來,我先挑幾個有特點的新模型,直接給你看效果,再詳細介紹前面那個語音輸入法。然後是這次很重要的一個能力,GGUF量化,最後統一說一下整合包怎麼升級,以及模型下載的一些細節。上次沒講到的地方,這次一起補上。先看這次新增的模型,它們主要集中在TTS和ASR兩個標线下。TTS裡最有特點的是,Index TTS2和Supertonic3。Index TTS2,它支持從文本內容推斷語氣的聲音合成。高級參數裡,寫上情感參考文本,模型會根據這段文本來推斷情緒,然後合成語音。比如這段文本,合成出來的語音大概率會帶憤怒和失望的語氣。請把桌上的文件整理好,下午3點前交給我。我再把情感強度調成0.8,重新合成,語氣聽上去就會更強烈。請把桌上的文件整理好,下午3點前交給我。如果去掉情感參考文本,勾選從朗讀文本推斷。那合成出來的就是這種效果。請把桌上的文件整理好,下午3點前交給我。需要注意的是,Index TTS2比較吃顯存。顯卡配置低的話,就用MOS系列,或者上期介紹的千問3 TTS0.6b。Supertonic3的特點是速度快,而且支持流式生存。生存模式切換到流式。第一次生存會重載後台服務,會多出一段加載時間。第二次就快了,所以建議第一次用一小段文本來預熱一下。看,我明白你什麼想的。他再次回到,男人在金色衣服,男人在走動的心靈跟一路復動的問題,正在做一場文件,像是一場文件的議題。流式的好處主要體現在長文本合成上,可以邊合成邊輸出語音。不用像離線模式那樣,要等所有文本合成結束,才返回完整的語音。需要注意的是,這個模型不支持中文。中文的流式合成現在只有Vox CP-M2支持,不過目前效果還在優化當中。但首次延遲高,流式生存,則邊接受文字,邊分段合成。中間有卡頓,想使用中文流式可以再等等。TTS說完,再看語音轉寫。這次也新增了五個模型。先說結論,中文轉寫推薦用千問3ASR 1.7B,準確率相當高,顯存夠的話就用Vib Voice ASR,能力更強,不精準還支持多人對話。要速度就用Limotron 3.5ASR 0.6B,但它中文識別的準確率要差一些。這是同一段音頻三個模型的識別對比,我把他們的結果丟給AI打分。最終判定千問3ASR 1.7B,轉寫的最準確。Limotron 3.5ASR 0.6B,同時支持離線和流式,勾選最下面的流式轉寫,它就會邊識別邊吐字。這種模式非常適合語音輸入,實施對話這些場景,不用等音頻全部識別完才返回結果。我本地的語音輸入就是拿它做的。具體實施過程其實不複雜。我先寫了一份大致需求,描述下功能要求,然後交給Codex。因為服務端是現成的,直接調用AudioCPP的服務。代碼只需要做語音輸入發送,顯示這一層。大概半小時就完成了,後面調適用了10分鐘左右。具體用法,先啟動Run Server ASR Stream服務,再運行Speak Type。按住Ctrl加Shift加空格,打開麥克風。然後把光標放進你的文本編輯軟件裡。記事本,Obsidian都可以,對著麥克風說話,內容就直接轉寫過來了。這次升級真正重要的,不是模型又多了幾個。如果想換行,期間你說完一句打個回車。需要說明一點,這程序只是個DEMO,主要是為了驗證用AudioCPP,開發落地應用的可能性。在它上面可以開發出WebUI界面,可以做實施語音對話,可以接OpenAI風格的API,往下還有剛才的語音輸入,瀏覽器擴展以及其他桌面應用等等,都可以做。現在回過頭,把這次介面上的幾個改動過一遍。右上角語言,現在支持繁體和英文,切換時大概有一到兩秒的延時。另一個新增功能是內置音頻管理。上傳參考音頻,寫好對應的文本,然後點保存,就會自動存到內置音頻列表裡。不用的可以選中刪除。界面上這次最大的功能更新,是模型的量化管理。點開GGUF,下面提示可轉換的,就能進行量化。先選幾位量化,然後點轉換按鈕。量化的好處是顯存佔用更小。對於沒有顯卡的機器,CPU運行量化板速度也更快。比如這個VibeOS ASR模型,原始模型加載後,大概要占16G顯存。8G顯存根本就跑不動。但把它轉成Q4量化板以後,加載就沒問題了。但是大參數模型轉換比較慢,像我這次用了20多分鐘。當有GGUF時,點這裡加載模型,後台服務會優先加載量化板。需要提醒一點,目前AudioCPP並不支持其他框架轉換的GGUF,並且也只有部分模型現在可以轉換。在列表裡選模型時,能不能轉換下面都有提示。轉好的量化板不想用也可以直接轉換。就能恢復原始權重。目前AudioCPP已經完整支持30個語音模型。所有模型的能力特點,我都整理成了配套筆記。可以通過這張表格來數查。鏈接我放在評論區置頂。最後說下整合包怎麼升級,以及模型下載。第一個鏈接是新的完整版整合包。如果你之前沒安裝過,就下這個。上期安裝過的,點第二個鏈接下載升級包。解壓以後,複製裡面的所有文件,覆蓋到原來整合包目錄下。然後全部替換,就升級好了。以後更新,可以直接點update。程序會自動檢測,有新版本會自動下載。雙擊Run WebUI打開界面。這個啟動窗口第一次要等幾十秒。程序加載完會自動打開網頁。再說下模型下載。目前AudioCPP的模型都要從Hugging Face上下載。並且不支持手動放模型。國內網絡如果下載有問題,可以試一下這裡填上代理地址。另外有個別模型,像Stable Audio這個系列,需要在Hugging Face上申請。具體操作步驟是這樣。先註册一個免費的Hugging Face帳號,然後到這個地址填寫申請表單。申請通過後,在典禮的Hugging Face頭像,菜單底部訪問Token。創建一個新Token,選紙讀。複製創建的Token,粘貼到WebUI的Token框裡,就可以下載模型了。模型下載機制,是先在Modus目錄下,創建一個臨時下載目錄,就是這個文件夾。既然能看到正在下載的模型。全部下載完成後,程序會自動創建重命名模型文件夾。下載過程中中斷也沒關係。繼續點下載按鈕,會斷點續傳。下載完成後點刷新,就可以下載模型了。程序也會自检模型文件的完整性。缺文件會提示模型不完整。新商手的朋友,想了解介面基礎操作,以及其他常用模型,可以看語音AI系統化的那期視頻。本期視頻的配套筆記,鏈接我放在評論區置頂。整個包使用中的問題,或者好的點子想法,都可以在評論區留言。我會定期收集整理,更新版本。這裡是AI探索與發現,感謝觀看,下期見。我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,我們在評論區,", "segments": [{"id": 0, "seek": 0, "start": 0.0, "end": 3.0, "text": "請評估本項目加入中英文支持。", "tokens": [50365, 16302, 39115, 7384, 108, 8802, 8313, 227, 11386, 43878, 5975, 27869, 17174, 35488, 1543, 50515], "temperature": 0, "avg_logprob": -0.18895156784813003, "compression_ratio": 1.045267489711934, "no_speech_prob": 6.9576636119172974e-12}, {"id": 1, "seek": 0, "start": 8.0, "end": 22.0, "text": "上期視頻我介紹了Audio CPP,當時我把它比作音頻版的Nama CPP,僅僅過了幾天,它又一次大升級,不僅一口氣新增了10個模型,還支持加載GGUF量化版,並且落地了牛市生成。", "tokens": [50765, 5708, 16786, 27333, 29940, 1654, 30312, 38193, 2289, 15591, 1004, 383, 17755, 11, 13118, 6611, 1654, 42061, 11706, 11914, 18034, 29940, 42096, 1546, 45, 2404, 383, 17755, 11, 9502, 227, 9502, 227, 8816, 2289, 23575, 6135, 11, 11284, 17047, 27505, 3582, 41670, 35315, 11, 1960, 9502, 227, 2257, 18144, 24090, 12560, 24228, 252, 2289, 3279, 3338, 41908, 39823, 11, 7824, 35488, 9990, 12290, 231, 27561, 52, 37, 26748, 23756, 42096, 11, 35934, 20334, 30848, 10928, 2289, 40003, 27261, 8244, 11336, 1543, 51465], "temperature": 0, "avg_logprob": -0.18895156784813003, "compression_ratio": 1.045267489711934, "no_speech_prob": 6.9576636119172974e-12}, {"id": 2, "seek": 2200, "start": 22.0, "end": 31.5, "text": "甚至現在,你看到的這個語音輸入法,就是用它做的。", "tokens": [50365, 17556, 20844, 12648, 11, 16529, 4511, 1546, 6287, 31348, 18034, 12290, 116, 14028, 11148, 11, 5620, 9254, 11284, 10907, 1546, 1543, 50840], "temperature": 0, "avg_logprob": -0.12271662875338718, "compression_ratio": 1.2064777327935223, "no_speech_prob": 8.911790576326961e-12}, {"id": 3, "seek": 2200, "start": 32.5, "end": 40.18, "text": "操作介面上,基礎用法沒變,主要改動是增加了多語言支持,GGUF量化轉換,還有音色庫管理。", "tokens": [50890, 13167, 235, 11914, 30312, 8833, 5708, 11, 26008, 14971, 236, 9254, 11148, 4913, 26327, 11, 13557, 4275, 34490, 12572, 1541, 24228, 252, 9990, 2289, 6392, 31348, 12009, 35488, 11, 27561, 52, 37, 26748, 23756, 32739, 36338, 11, 15569, 18034, 17673, 6346, 104, 23131, 13876, 1543, 51274], "temperature": 0, "avg_logprob": -0.12271662875338718, "compression_ratio": 1.2064777327935223, "no_speech_prob": 8.911790576326961e-12}, {"id": 4, "seek": 2200, "start": 40.7, "end": 46.480000000000004, "text": "接下來,我先挑幾個有特點的新模型,直接給你看效果,再詳細介紹前面那個語音輸入法。", "tokens": [51300, 40012, 11, 1654, 10108, 49067, 23575, 3338, 2412, 17682, 8216, 1546, 12560, 41908, 39823, 11, 43297, 17798, 16529, 43076, 9319, 11, 8623, 6716, 111, 41588, 30312, 38193, 8945, 8833, 20754, 31348, 18034, 12290, 116, 14028, 11148, 1543, 51589], "temperature": 0, "avg_logprob": -0.12271662875338718, "compression_ratio": 1.2064777327935223, "no_speech_prob": 8.911790576326961e-12}, {"id": 5, "seek": 4648, "start": 46.48, "end": 54.48, "text": "然後是這次很重要的一個能力,GGUF量化,最後統一說一下整合包怎麼升級,以及模型下載的一些細節。", "tokens": [50365, 10213, 1541, 2664, 9487, 4563, 24928, 1546, 8990, 8225, 13486, 11, 27561, 52, 37, 26748, 23756, 11, 20578, 33725, 2257, 4622, 8861, 27662, 14245, 23305, 11614, 41670, 35315, 11, 40282, 41908, 39823, 4438, 12290, 231, 1546, 38515, 41588, 27694, 1543, 50765], "temperature": 0, "avg_logprob": -0.13740269797188895, "compression_ratio": 1.1864951768488745, "no_speech_prob": 1.544589109392369e-11}, {"id": 6, "seek": 4648, "start": 54.48, "end": 56.98, "text": "上次沒講到的地方,這次一起補上。", "tokens": [50765, 5708, 9487, 4913, 11932, 4511, 1546, 30146, 11, 2664, 9487, 29567, 8083, 250, 5708, 1543, 50890], "temperature": 0, "avg_logprob": -0.13740269797188895, "compression_ratio": 1.1864951768488745, "no_speech_prob": 1.544589109392369e-11}, {"id": 7, "seek": 4648, "start": 56.98, "end": 61.98, "text": "先看這次新增的模型,它們主要集中在TTS和ASR兩個標线下。", "tokens": [50890, 10108, 4200, 2664, 9487, 12560, 24228, 252, 1546, 41908, 39823, 11, 11284, 4623, 13557, 4275, 26020, 5975, 3581, 51, 7327, 12565, 3160, 49, 34623, 46299, 16853, 123, 4438, 1543, 51140], "temperature": 0, "avg_logprob": -0.13740269797188895, "compression_ratio": 1.1864951768488745, "no_speech_prob": 1.544589109392369e-11}, {"id": 8, "seek": 4648, "start": 61.98, "end": 66.97999999999999, "text": "TTS裡最有特點的是,Index TTS2和Supertonic3。", "tokens": [51140, 51, 7327, 11066, 8661, 2412, 17682, 8216, 24620, 11, 21790, 3121, 314, 7327, 17, 12565, 30152, 1756, 299, 18, 1543, 51390], "temperature": 0, "avg_logprob": -0.13740269797188895, "compression_ratio": 1.1864951768488745, "no_speech_prob": 1.544589109392369e-11}, {"id": 9, "seek": 4648, "start": 66.97999999999999, "end": 71.47999999999999, "text": "Index TTS2,它支持從文本內容推斷語氣的聲音合成。", "tokens": [51390, 21790, 3121, 314, 7327, 17, 11, 11284, 35488, 21068, 17174, 8802, 28472, 25750, 33597, 4307, 115, 31348, 24090, 1546, 29454, 18034, 14245, 11336, 1543, 51615], "temperature": 0, "avg_logprob": -0.13740269797188895, "compression_ratio": 1.1864951768488745, "no_speech_prob": 1.544589109392369e-11}, {"id": 10, "seek": 7148, "start": 71.48, "end": 78.48, "text": "高級參數裡,寫上情感參考文本,模型會根據這段文本來推斷情緒,然後合成語音。", "tokens": [50365, 12979, 35315, 47375, 30622, 11066, 11, 4510, 104, 5708, 10570, 9709, 47375, 26504, 17174, 8802, 11, 41908, 39823, 6236, 31337, 36841, 2664, 28427, 17174, 8802, 3763, 33597, 4307, 115, 10570, 12608, 240, 11, 10213, 14245, 11336, 31348, 18034, 1543, 50715], "temperature": 0, "avg_logprob": -0.06348626565613202, "compression_ratio": 1.4803149606299213, "no_speech_prob": 2.0687163787558305e-11}, {"id": 11, "seek": 7148, "start": 78.48, "end": 83.48, "text": "比如這段文本,合成出來的語音大概率會帶憤怒和失望的語氣。", "tokens": [50715, 36757, 2664, 28427, 17174, 8802, 11, 14245, 11336, 29741, 1546, 31348, 18034, 32044, 44866, 6236, 32379, 29713, 97, 3757, 240, 12565, 28564, 22694, 1546, 31348, 24090, 1543, 50965], "temperature": 0, "avg_logprob": -0.06348626565613202, "compression_ratio": 1.4803149606299213, "no_speech_prob": 2.0687163787558305e-11}, {"id": 12, "seek": 7148, "start": 83.48, "end": 87.48, "text": "請把桌上的文件整理好,下午3點前交給我。", "tokens": [50965, 16302, 16075, 162, 5475, 5708, 1546, 17174, 20485, 27662, 13876, 2131, 11, 4438, 44237, 18, 8216, 8945, 28455, 17798, 1654, 1543, 51165], "temperature": 0, "avg_logprob": -0.06348626565613202, "compression_ratio": 1.4803149606299213, "no_speech_prob": 2.0687163787558305e-11}, {"id": 13, "seek": 7148, "start": 87.48, "end": 92.48, "text": "我再把情感強度調成0.8,重新合成,語氣聽上去就會更強烈。", "tokens": [51165, 1654, 8623, 16075, 10570, 9709, 22752, 13127, 32943, 11336, 15, 13, 23, 11, 12624, 12560, 14245, 11336, 11, 31348, 24090, 21524, 5708, 6734, 47778, 19002, 22752, 23661, 230, 1543, 51415], "temperature": 0, "avg_logprob": -0.06348626565613202, "compression_ratio": 1.4803149606299213, "no_speech_prob": 2.0687163787558305e-11}, {"id": 14, "seek": 7148, "start": 92.48, "end": 96.48, "text": "請把桌上的文件整理好,下午3點前交給我。", "tokens": [51415, 16302, 16075, 162, 5475, 5708, 1546, 17174, 20485, 27662, 13876, 2131, 11, 4438, 44237, 18, 8216, 8945, 28455, 17798, 1654, 1543, 51615], "temperature": 0, "avg_logprob": -0.06348626565613202, "compression_ratio": 1.4803149606299213, "no_speech_prob": 2.0687163787558305e-11}, {"id": 15, "seek": 9648, "start": 96.48, "end": 101.48, "text": "如果去掉情感參考文本,勾選從朗讀文本推斷。", "tokens": [50365, 13119, 6734, 29327, 10570, 9709, 47375, 26504, 17174, 8802, 11, 7978, 122, 21625, 21068, 1474, 245, 7422, 222, 17174, 8802, 33597, 4307, 115, 1543, 50615], "temperature": 0, "avg_logprob": -0.08085579342312282, "compression_ratio": 1.1635220125786163, "no_speech_prob": 1.863473264585025e-11}, {"id": 16, "seek": 9648, "start": 101.48, "end": 103.48, "text": "那合成出來的就是這種效果。", "tokens": [50615, 4184, 14245, 11336, 29741, 1546, 5620, 33005, 43076, 9319, 1543, 50715], "temperature": 0, "avg_logprob": -0.08085579342312282, "compression_ratio": 1.1635220125786163, "no_speech_prob": 1.863473264585025e-11}, {"id": 17, "seek": 9648, "start": 103.48, "end": 108.48, "text": "請把桌上的文件整理好,下午3點前交給我。", "tokens": [50715, 16302, 16075, 162, 5475, 5708, 1546, 17174, 20485, 27662, 13876, 2131, 11, 4438, 44237, 18, 8216, 8945, 28455, 17798, 1654, 1543, 50965], "temperature": 0, "avg_logprob": -0.08085579342312282, "compression_ratio": 1.1635220125786163, "no_speech_prob": 1.863473264585025e-11}, {"id": 18, "seek": 9648, "start": 108.48, "end": 111.48, "text": "需要注意的是,Index TTS2比較吃顯存。", "tokens": [50965, 35748, 40794, 24620, 11, 21790, 3121, 314, 7327, 17, 25174, 10123, 46605, 39781, 1543, 51115], "temperature": 0, "avg_logprob": -0.08085579342312282, "compression_ratio": 1.1635220125786163, "no_speech_prob": 1.863473264585025e-11}, {"id": 19, "seek": 9648, "start": 111.48, "end": 117.48, "text": "顯卡配置低的話,就用MOS系列,或者上期介紹的千問3 TTS0.6b。", "tokens": [51115, 46605, 32681, 38846, 34719, 41377, 21358, 11, 3111, 9254, 44, 4367, 25368, 43338, 11, 31148, 5708, 16786, 30312, 38193, 1546, 20787, 11361, 18, 314, 7327, 15, 13, 21, 65, 1543, 51415], "temperature": 0, "avg_logprob": -0.08085579342312282, "compression_ratio": 1.1635220125786163, "no_speech_prob": 1.863473264585025e-11}, {"id": 20, "seek": 9648, "start": 117.48, "end": 121.48, "text": "Supertonic3的特點是速度快,而且支持流式生存。", "tokens": [51415, 30152, 1756, 299, 18, 1546, 17682, 8216, 1541, 31217, 13127, 10251, 11, 22942, 35488, 27854, 27584, 8244, 39781, 1543, 51615], "temperature": 0, "avg_logprob": -0.08085579342312282, "compression_ratio": 1.1635220125786163, "no_speech_prob": 1.863473264585025e-11}, {"id": 21, "seek": 9648, "start": 121.48, "end": 123.48, "text": "生存模式切換到流式。", "tokens": [51615, 8244, 39781, 41908, 27584, 23632, 36338, 4511, 27854, 27584, 1543, 51715], "temperature": 0, "avg_logprob": -0.08085579342312282, "compression_ratio": 1.1635220125786163, "no_speech_prob": 1.863473264585025e-11}, {"id": 22, "seek": 12348, "start": 123.48, "end": 127.48, "text": "第一次生存會重載後台服務,會多出一段加載時間。", "tokens": [50365, 18049, 9487, 8244, 39781, 6236, 12624, 12290, 231, 5661, 15433, 27408, 39517, 11, 6236, 6392, 7781, 2257, 28427, 9990, 12290, 231, 20788, 1543, 50565], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 23, "seek": 12348, "start": 127.48, "end": 131.48000000000002, "text": "第二次就快了,所以建議第一次用一小段文本來預熱一下。", "tokens": [50565, 19693, 9487, 3111, 10251, 2289, 11, 7239, 34157, 24686, 18049, 9487, 9254, 2257, 7322, 28427, 17174, 8802, 3763, 30156, 32745, 8861, 1543, 50765], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 24, "seek": 12348, "start": 131.48000000000002, "end": 132.48000000000002, "text": "看,我明白你什麼想的。", "tokens": [50765, 4200, 11, 1654, 28430, 2166, 7598, 7093, 1546, 1543, 50815], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 25, "seek": 12348, "start": 132.48000000000002, "end": 134.48000000000002, "text": "他再次回到,", "tokens": [50815, 5000, 8623, 9487, 8350, 4511, 11, 50915], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 26, "seek": 12348, "start": 134.48000000000002, "end": 135.48000000000002, "text": "男人在金色衣服,", "tokens": [50915, 25243, 4035, 3581, 19117, 17673, 9890, 96, 27408, 11, 50965], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 27, "seek": 12348, "start": 135.48000000000002, "end": 137.48000000000002, "text": "男人在走動的心靈跟一路復動的問題,", "tokens": [50965, 25243, 4035, 3581, 9575, 12572, 1546, 7945, 5363, 230, 9678, 2257, 24658, 2172, 102, 12572, 1546, 17197, 11, 51065], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 28, "seek": 12348, "start": 137.48000000000002, "end": 140.48000000000002, "text": "正在做一場文件,像是一場文件的議題。", "tokens": [51065, 15789, 3581, 10907, 2257, 21446, 17174, 20485, 11, 12760, 1541, 2257, 21446, 17174, 20485, 1546, 24686, 14475, 1543, 51215], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 29, "seek": 12348, "start": 140.48000000000002, "end": 143.48000000000002, "text": "流式的好處主要體現在長文本合成上,", "tokens": [51215, 27854, 27584, 1546, 2131, 39289, 13557, 4275, 23987, 12648, 15353, 17174, 8802, 14245, 11336, 5708, 11, 51365], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 30, "seek": 12348, "start": 143.48000000000002, "end": 145.48000000000002, "text": "可以邊合成邊輸出語音。", "tokens": [51365, 6723, 14901, 14245, 11336, 14901, 12290, 116, 7781, 31348, 18034, 1543, 51465], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 31, "seek": 12348, "start": 145.48000000000002, "end": 148.48000000000002, "text": "不用像離線模式那樣,要等所有文本合成結束,", "tokens": [51465, 24384, 12760, 32662, 31699, 41908, 27584, 4184, 6180, 11, 4275, 10187, 39300, 17174, 8802, 14245, 11336, 17144, 44320, 11, 51615], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 32, "seek": 12348, "start": 148.48000000000002, "end": 150.48000000000002, "text": "才返回完整的語音。", "tokens": [51615, 18888, 3316, 242, 8350, 14128, 27662, 1546, 31348, 18034, 1543, 51715], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 33, "seek": 12348, "start": 150.48000000000002, "end": 152.48000000000002, "text": "需要注意的是,這個模型不支持中文。", "tokens": [51715, 35748, 40794, 24620, 11, 6287, 41908, 39823, 1960, 35488, 5975, 17174, 1543, 51815], "temperature": 0, "avg_logprob": -0.5147750317750863, "compression_ratio": 1.4285714285714286, "no_speech_prob": 2.2361810320115083e-11}, {"id": 34, "seek": 15248, "start": 152.48, "end": 156.48, "text": "中文的流式合成現在只有Vox CP-M2支持,", "tokens": [50365, 5975, 17174, 1546, 27854, 27584, 14245, 11336, 12648, 35244, 53, 5230, 22431, 12, 44, 17, 35488, 11, 50565], "temperature": 0, "avg_logprob": -0.11462981182595958, "compression_ratio": 1.126530612244898, "no_speech_prob": 1.7538609453637832e-11}, {"id": 35, "seek": 15248, "start": 156.48, "end": 158.48, "text": "不過目前效果還在優化當中。", "tokens": [50565, 28193, 39004, 43076, 9319, 7824, 3581, 12827, 103, 23756, 13118, 5975, 1543, 50665], "temperature": 0, "avg_logprob": -0.11462981182595958, "compression_ratio": 1.126530612244898, "no_speech_prob": 1.7538609453637832e-11}, {"id": 36, "seek": 15248, "start": 158.48, "end": 161.48, "text": "但首次延遲高,流式生存,", "tokens": [50665, 8395, 25444, 9487, 17527, 114, 3330, 110, 12979, 11, 27854, 27584, 8244, 39781, 11, 50815], "temperature": 0, "avg_logprob": -0.11462981182595958, "compression_ratio": 1.126530612244898, "no_speech_prob": 1.7538609453637832e-11}, {"id": 37, "seek": 15248, "start": 161.48, "end": 164.48, "text": "則邊接受文字,邊分段合成。", "tokens": [50815, 46225, 14901, 14468, 23151, 17174, 22381, 11, 14901, 6627, 28427, 14245, 11336, 1543, 50965], "temperature": 0, "avg_logprob": -0.11462981182595958, "compression_ratio": 1.126530612244898, "no_speech_prob": 1.7538609453637832e-11}, {"id": 38, "seek": 15248, "start": 164.48, "end": 167.48, "text": "中間有卡頓,想使用中文流式可以再等等。", "tokens": [50965, 5975, 11016, 2412, 32681, 8313, 241, 11, 7093, 22982, 9254, 5975, 17174, 27854, 27584, 6723, 8623, 36000, 1543, 51115], "temperature": 0, "avg_logprob": -0.11462981182595958, "compression_ratio": 1.126530612244898, "no_speech_prob": 1.7538609453637832e-11}, {"id": 39, "seek": 15248, "start": 167.48, "end": 170.48, "text": "TTS說完,再看語音轉寫。", "tokens": [51115, 51, 7327, 4622, 14128, 11, 8623, 4200, 31348, 18034, 32739, 4510, 104, 1543, 51265], "temperature": 0, "avg_logprob": -0.11462981182595958, "compression_ratio": 1.126530612244898, "no_speech_prob": 1.7538609453637832e-11}, {"id": 40, "seek": 15248, "start": 170.48, "end": 172.48, "text": "這次也新增了五個模型。", "tokens": [51265, 2664, 9487, 6404, 12560, 24228, 252, 2289, 21001, 3338, 41908, 39823, 1543, 51365], "temperature": 0, "avg_logprob": -0.11462981182595958, "compression_ratio": 1.126530612244898, "no_speech_prob": 1.7538609453637832e-11}, {"id": 41, "seek": 17248, "start": 172.48, "end": 176.48, "text": "先說結論,中文轉寫推薦用千問3ASR 1.7B,", "tokens": [50365, 10108, 4622, 17144, 29340, 11, 5975, 17174, 32739, 4510, 104, 33597, 27746, 99, 9254, 20787, 11361, 18, 3160, 49, 502, 13, 22, 33, 11, 50565], "temperature": 0, "avg_logprob": -0.20142088584529544, "compression_ratio": 1.0405405405405406, "no_speech_prob": 2.0135467945769925e-11}, {"id": 42, "seek": 17248, "start": 176.48, "end": 180.48, "text": "準確率相當高,顯存夠的話就用Vib Voice ASR,", "tokens": [50565, 28220, 24293, 44866, 15106, 13118, 12979, 11, 46605, 39781, 31649, 21358, 3111, 9254, 53, 897, 15229, 7469, 49, 11, 50765], "temperature": 0, "avg_logprob": -0.20142088584529544, "compression_ratio": 1.0405405405405406, "no_speech_prob": 2.0135467945769925e-11}, {"id": 43, "seek": 17248, "start": 180.48, "end": 183.48, "text": "能力更強,不精準還支持多人對話。", "tokens": [50765, 8225, 13486, 19002, 22752, 11, 1960, 34910, 28220, 7824, 35488, 6392, 4035, 2855, 11103, 1543, 50915], "temperature": 0, "avg_logprob": -0.20142088584529544, "compression_ratio": 1.0405405405405406, "no_speech_prob": 2.0135467945769925e-11}, {"id": 44, "seek": 17248, "start": 183.48, "end": 186.48, "text": "要速度就用Limotron 3.5ASR 0.6B,", "tokens": [50915, 4275, 31217, 13127, 3111, 9254, 43, 332, 310, 2044, 805, 13, 20, 3160, 49, 1958, 13, 21, 33, 11, 51065], "temperature": 0, "avg_logprob": -0.20142088584529544, "compression_ratio": 1.0405405405405406, "no_speech_prob": 2.0135467945769925e-11}, {"id": 45, "seek": 17248, "start": 186.48, "end": 189.48, "text": "但它中文識別的準確率要差一些。", "tokens": [51065, 8395, 11284, 5975, 17174, 43143, 16158, 1546, 28220, 24293, 44866, 4275, 21679, 38515, 1543, 51215], "temperature": 0, "avg_logprob": -0.20142088584529544, "compression_ratio": 1.0405405405405406, "no_speech_prob": 2.0135467945769925e-11}, {"id": 46, "seek": 18948, "start": 189.48, "end": 192.48, "text": "這是同一段音頻三個模型的識別對比,", "tokens": [50365, 16600, 13089, 2257, 28427, 18034, 29940, 10960, 3338, 41908, 39823, 1546, 43143, 16158, 2855, 11706, 11, 50515], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 47, "seek": 18948, "start": 192.48, "end": 194.48, "text": "我把他們的結果丟給AI打分。", "tokens": [50515, 1654, 16075, 20486, 1546, 35181, 940, 253, 17798, 48698, 12467, 6627, 1543, 50615], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 48, "seek": 18948, "start": 194.48, "end": 197.48, "text": "最終判定千問3ASR 1.7B,", "tokens": [50615, 8661, 29371, 2437, 97, 12088, 20787, 11361, 18, 3160, 49, 502, 13, 22, 33, 11, 50765], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 49, "seek": 18948, "start": 197.48, "end": 198.48, "text": "轉寫的最準確。", "tokens": [50765, 32739, 4510, 104, 1546, 8661, 28220, 24293, 1543, 50815], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 50, "seek": 18948, "start": 198.48, "end": 201.48, "text": "Limotron 3.5ASR 0.6B,", "tokens": [50815, 43, 332, 310, 2044, 805, 13, 20, 3160, 49, 1958, 13, 21, 33, 11, 50965], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 51, "seek": 18948, "start": 201.48, "end": 203.48, "text": "同時支持離線和流式,", "tokens": [50965, 13089, 6611, 35488, 32662, 31699, 12565, 27854, 27584, 11, 51065], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 52, "seek": 18948, "start": 203.48, "end": 205.48, "text": "勾選最下面的流式轉寫,", "tokens": [51065, 7978, 122, 21625, 8661, 47150, 1546, 27854, 27584, 32739, 4510, 104, 11, 51165], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 53, "seek": 18948, "start": 205.48, "end": 207.48, "text": "它就會邊識別邊吐字。", "tokens": [51165, 11284, 47778, 14901, 43143, 16158, 14901, 2392, 238, 22381, 1543, 51265], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 54, "seek": 18948, "start": 207.48, "end": 209.48, "text": "這種模式非常適合語音輸入,", "tokens": [51265, 33005, 41908, 27584, 14392, 3330, 102, 14245, 31348, 18034, 12290, 116, 14028, 11, 51365], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 55, "seek": 18948, "start": 209.48, "end": 211.48, "text": "實施對話這些場景,", "tokens": [51365, 10376, 4307, 121, 2855, 11103, 29869, 21446, 50218, 11, 51465], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 56, "seek": 18948, "start": 211.48, "end": 213.48, "text": "不用等音頻全部識別完才返回結果。", "tokens": [51465, 24384, 10187, 18034, 29940, 38714, 43143, 16158, 14128, 18888, 3316, 242, 8350, 35181, 1543, 51565], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 57, "seek": 18948, "start": 213.48, "end": 216.48, "text": "我本地的語音輸入就是拿它做的。", "tokens": [51565, 1654, 8802, 10928, 1546, 31348, 18034, 12290, 116, 14028, 5620, 24351, 11284, 10907, 1546, 1543, 51715], "temperature": 0, "avg_logprob": -0.08584722066890298, "compression_ratio": 1.2078313253012047, "no_speech_prob": 2.1786674894719305e-11}, {"id": 58, "seek": 21648, "start": 216.48, "end": 219.48, "text": "具體實施過程其實不複雜。", "tokens": [50365, 39806, 23987, 10376, 4307, 121, 8816, 29649, 14139, 1960, 164, 97, 229, 6306, 250, 1543, 50515], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 59, "seek": 21648, "start": 219.48, "end": 221.48, "text": "我先寫了一份大致需求,", "tokens": [50515, 1654, 10108, 4510, 104, 2289, 2257, 36266, 3582, 6784, 112, 32535, 32718, 11, 50615], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 60, "seek": 21648, "start": 221.48, "end": 223.48, "text": "描述下功能要求,", "tokens": [50615, 11673, 237, 3316, 108, 4438, 32311, 8225, 4275, 32718, 11, 50715], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 61, "seek": 21648, "start": 223.48, "end": 224.48, "text": "然後交給Codex。", "tokens": [50715, 10213, 28455, 17798, 34, 1429, 87, 1543, 50765], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 62, "seek": 21648, "start": 224.48, "end": 226.48, "text": "因為服務端是現成的,", "tokens": [50765, 11471, 27408, 39517, 11957, 107, 1541, 9581, 11336, 1546, 11, 50865], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 63, "seek": 21648, "start": 226.48, "end": 228.48, "text": "直接調用AudioCPP的服務。", "tokens": [50865, 43297, 32943, 9254, 15591, 1004, 34, 17755, 1546, 27408, 39517, 1543, 50965], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 64, "seek": 21648, "start": 228.48, "end": 230.48, "text": "代碼只需要做語音輸入發送,", "tokens": [50965, 19105, 16337, 120, 14003, 35748, 10907, 31348, 18034, 12290, 116, 14028, 14637, 29309, 11, 51065], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 65, "seek": 21648, "start": 230.48, "end": 231.48, "text": "顯示這一層。", "tokens": [51065, 46605, 25696, 32260, 9636, 97, 1543, 51115], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 66, "seek": 21648, "start": 231.48, "end": 233.48, "text": "大概半小時就完成了,", "tokens": [51115, 32044, 30018, 46997, 3111, 41509, 2289, 11, 51215], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 67, "seek": 21648, "start": 233.48, "end": 235.48, "text": "後面調適用了10分鐘左右。", "tokens": [51215, 5661, 8833, 32943, 3330, 102, 9254, 2289, 3279, 32611, 29457, 1543, 51315], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 68, "seek": 21648, "start": 235.48, "end": 237.48, "text": "具體用法,", "tokens": [51315, 39806, 23987, 9254, 11148, 11, 51415], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 69, "seek": 21648, "start": 237.48, "end": 240.48, "text": "先啟動Run Server ASR Stream服務,", "tokens": [51415, 10108, 3284, 253, 12572, 42639, 25684, 7469, 49, 24904, 27408, 39517, 11, 51565], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 70, "seek": 21648, "start": 240.48, "end": 242.48, "text": "再運行Speak Type。", "tokens": [51565, 8623, 32772, 8082, 19588, 514, 15576, 1543, 51665], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 71, "seek": 21648, "start": 242.48, "end": 245.48, "text": "按住Ctrl加Shift加空格,", "tokens": [51665, 26613, 21632, 34, 28269, 9990, 7774, 2008, 9990, 23322, 30921, 11, 51815], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 72, "seek": 21648, "start": 245.48, "end": 246.48, "text": "打開麥克風。", "tokens": [51815, 12467, 8949, 3994, 98, 24881, 22713, 1543, 51865], "temperature": 0, "avg_logprob": -0.12245535985224665, "compression_ratio": 1.1376404494382022, "no_speech_prob": 1.814022022872397e-11}, {"id": 73, "seek": 24648, "start": 246.48, "end": 248.48, "text": "然後把光標放進你的文本編輯軟件裡。", "tokens": [50365, 10213, 16075, 20690, 46299, 12744, 18214, 18961, 17174, 8802, 12608, 101, 12290, 107, 14454, 253, 20485, 11066, 1543, 50465], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 74, "seek": 24648, "start": 248.48, "end": 249.48, "text": "記事本,", "tokens": [50465, 16958, 6973, 8802, 11, 50515], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 75, "seek": 24648, "start": 249.48, "end": 250.48, "text": "Obsidian都可以,", "tokens": [50515, 46, 929, 34681, 7182, 6723, 11, 50565], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 76, "seek": 24648, "start": 250.48, "end": 251.48, "text": "對著麥克風說話,", "tokens": [50565, 2855, 19382, 3994, 98, 24881, 22713, 4622, 11103, 11, 50615], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 77, "seek": 24648, "start": 251.48, "end": 253.48, "text": "內容就直接轉寫過來了。", "tokens": [50615, 28472, 25750, 3111, 43297, 32739, 4510, 104, 8816, 25364, 1543, 50715], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 78, "seek": 24648, "start": 253.48, "end": 255.48, "text": "這次升級真正重要的,", "tokens": [50715, 2664, 9487, 41670, 35315, 6303, 15789, 24928, 1546, 11, 50815], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 79, "seek": 24648, "start": 255.48, "end": 258.48, "text": "不是模型又多了幾個。", "tokens": [50815, 7296, 41908, 39823, 17047, 6392, 2289, 23575, 3338, 1543, 50965], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 80, "seek": 24648, "start": 258.48, "end": 259.48, "text": "如果想換行,", "tokens": [50965, 13119, 7093, 36338, 8082, 11, 51015], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 81, "seek": 24648, "start": 259.48, "end": 261.48, "text": "期間你說完一句打個回車。", "tokens": [51015, 16786, 11016, 42920, 14128, 2257, 34592, 12467, 3338, 8350, 25941, 1543, 51115], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 82, "seek": 24648, "start": 261.48, "end": 262.48, "text": "需要說明一點,", "tokens": [51115, 35748, 4622, 11100, 19202, 11, 51165], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 83, "seek": 24648, "start": 262.48, "end": 264.48, "text": "這程序只是個DEMO,", "tokens": [51165, 2664, 29649, 6346, 237, 36859, 3338, 35, 6683, 46, 11, 51265], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 84, "seek": 24648, "start": 264.48, "end": 266.48, "text": "主要是為了驗證用AudioCPP,", "tokens": [51265, 13557, 4275, 1541, 6344, 2289, 24023, 245, 11823, 231, 9254, 15591, 1004, 34, 17755, 11, 51365], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 85, "seek": 24648, "start": 266.48, "end": 268.48, "text": "開發落地應用的可能性。", "tokens": [51365, 8949, 14637, 30848, 10928, 20481, 9254, 1546, 16657, 21686, 1543, 51465], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 86, "seek": 24648, "start": 268.48, "end": 271.48, "text": "在它上面可以開發出WebUI界面,", "tokens": [51465, 3581, 11284, 49750, 6723, 8949, 14637, 7781, 4360, 65, 46324, 21185, 8833, 11, 51615], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 87, "seek": 24648, "start": 271.48, "end": 273.48, "text": "可以做實施語音對話,", "tokens": [51615, 6723, 10907, 10376, 4307, 121, 31348, 18034, 2855, 11103, 11, 51715], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 88, "seek": 24648, "start": 273.48, "end": 275.48, "text": "可以接OpenAI風格的API,", "tokens": [51715, 6723, 14468, 45569, 48698, 22713, 30921, 1546, 4715, 40, 11, 51815], "temperature": 0, "avg_logprob": -0.09986525393546895, "compression_ratio": 1.2016129032258065, "no_speech_prob": 1.5091259839006277e-11}, {"id": 89, "seek": 27548, "start": 275.48, "end": 277.48, "text": "往下還有剛才的語音輸入,", "tokens": [50365, 29510, 4438, 15569, 16940, 18888, 1546, 31348, 18034, 12290, 116, 14028, 11, 50465], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 90, "seek": 27548, "start": 277.48, "end": 279.48, "text": "瀏覽器擴展以及其他桌面應用等等,", "tokens": [50465, 163, 222, 237, 2888, 121, 34386, 13167, 112, 43491, 40282, 9572, 5000, 162, 5475, 8833, 20481, 9254, 36000, 11, 50565], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 91, "seek": 27548, "start": 279.48, "end": 280.48, "text": "都可以做。", "tokens": [50565, 7182, 6723, 10907, 1543, 50615], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 92, "seek": 27548, "start": 280.48, "end": 281.48, "text": "現在回過頭,", "tokens": [50615, 12648, 8350, 8816, 19921, 11, 50665], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 93, "seek": 27548, "start": 281.48, "end": 284.48, "text": "把這次介面上的幾個改動過一遍。", "tokens": [50665, 16075, 2664, 9487, 30312, 8833, 5708, 1546, 23575, 3338, 34490, 12572, 8816, 2257, 3330, 235, 1543, 50815], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 94, "seek": 27548, "start": 284.48, "end": 285.48, "text": "右上角語言,", "tokens": [50815, 22060, 5708, 29389, 31348, 12009, 11, 50865], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 95, "seek": 27548, "start": 285.48, "end": 287.48, "text": "現在支持繁體和英文,", "tokens": [50865, 12648, 35488, 23141, 223, 23987, 12565, 27869, 17174, 11, 50965], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 96, "seek": 27548, "start": 287.48, "end": 289.48, "text": "切換時大概有一到兩秒的延時。", "tokens": [50965, 23632, 36338, 6611, 32044, 32241, 4511, 16313, 20956, 1546, 17527, 114, 6611, 1543, 51065], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 97, "seek": 27548, "start": 289.48, "end": 292.48, "text": "另一個新增功能是內置音頻管理。", "tokens": [51065, 22762, 8990, 12560, 24228, 252, 32311, 8225, 1541, 28472, 34719, 18034, 29940, 23131, 13876, 1543, 51215], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 98, "seek": 27548, "start": 292.48, "end": 294.48, "text": "上傳參考音頻,", "tokens": [51215, 5708, 40852, 47375, 26504, 18034, 29940, 11, 51315], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 99, "seek": 27548, "start": 294.48, "end": 295.48, "text": "寫好對應的文本,", "tokens": [51315, 4510, 104, 2131, 2855, 20481, 1546, 17174, 8802, 11, 51365], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 100, "seek": 27548, "start": 295.48, "end": 297.48, "text": "然後點保存,", "tokens": [51365, 10213, 8216, 24302, 39781, 11, 51465], "temperature": 0, "avg_logprob": -0.10654083016800554, "compression_ratio": 1.250909090909091, "no_speech_prob": 2.196027908130116e-11}, {"id": 101, "seek": 29748, "start": 297.48, "end": 299.48, "text": "就會自動存到內置音頻列表裡。", "tokens": [50365, 3111, 6236, 9722, 12572, 39781, 4511, 28472, 34719, 18034, 29940, 43338, 17571, 11066, 1543, 50465], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 102, "seek": 29748, "start": 299.48, "end": 301.48, "text": "不用的可以選中刪除。", "tokens": [50465, 24384, 1546, 6723, 21625, 5975, 2437, 103, 32999, 1543, 50565], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 103, "seek": 29748, "start": 301.48, "end": 304.48, "text": "界面上這次最大的功能更新,", "tokens": [50565, 21185, 8833, 5708, 2664, 9487, 8661, 39156, 32311, 8225, 19002, 12560, 11, 50715], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 104, "seek": 29748, "start": 304.48, "end": 305.48, "text": "是模型的量化管理。", "tokens": [50715, 1541, 41908, 39823, 1546, 26748, 23756, 23131, 13876, 1543, 50765], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 105, "seek": 29748, "start": 305.48, "end": 307.48, "text": "點開GGUF,", "tokens": [50765, 8216, 8949, 27561, 52, 37, 11, 50865], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 106, "seek": 29748, "start": 307.48, "end": 308.48, "text": "下面提示可轉換的,", "tokens": [50865, 47150, 20949, 25696, 4429, 32739, 36338, 1546, 11, 50915], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 107, "seek": 29748, "start": 308.48, "end": 310.48, "text": "就能進行量化。", "tokens": [50915, 3111, 8225, 18214, 8082, 26748, 23756, 1543, 51015], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 108, "seek": 29748, "start": 310.48, "end": 311.48, "text": "先選幾位量化,", "tokens": [51015, 10108, 21625, 23575, 11160, 26748, 23756, 11, 51065], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 109, "seek": 29748, "start": 311.48, "end": 313.48, "text": "然後點轉換按鈕。", "tokens": [51065, 10213, 8216, 32739, 36338, 26613, 41597, 243, 1543, 51165], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 110, "seek": 29748, "start": 313.48, "end": 315.48, "text": "量化的好處是顯存佔用更小。", "tokens": [51165, 26748, 23756, 1546, 2131, 39289, 1541, 46605, 39781, 1593, 242, 9254, 19002, 7322, 1543, 51265], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 111, "seek": 29748, "start": 315.48, "end": 317.48, "text": "對於沒有顯卡的機器,", "tokens": [51265, 2855, 19488, 6963, 46605, 32681, 1546, 17543, 34386, 11, 51365], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 112, "seek": 29748, "start": 317.48, "end": 319.48, "text": "CPU運行量化板速度也更快。", "tokens": [51365, 34, 8115, 32772, 8082, 26748, 23756, 43664, 31217, 13127, 6404, 19002, 10251, 1543, 51465], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 113, "seek": 29748, "start": 319.48, "end": 321.48, "text": "比如這個VibeOS ASR模型,", "tokens": [51465, 36757, 6287, 41093, 650, 4367, 7469, 49, 41908, 39823, 11, 51565], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 114, "seek": 29748, "start": 321.48, "end": 323.48, "text": "原始模型加載後,", "tokens": [51565, 19683, 15476, 41908, 39823, 9990, 12290, 231, 5661, 11, 51665], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 115, "seek": 29748, "start": 323.48, "end": 324.48, "text": "大概要占16G顯存。", "tokens": [51665, 32044, 4275, 5322, 254, 6866, 38, 46605, 39781, 1543, 51715], "temperature": 0, "avg_logprob": -0.12076969580216841, "compression_ratio": 1.2432432432432432, "no_speech_prob": 2.0314814067057263e-11}, {"id": 116, "seek": 32448, "start": 324.48, "end": 326.48, "text": "8G顯存根本就跑不動。", "tokens": [50365, 23, 38, 46605, 39781, 31337, 8802, 3111, 32585, 1960, 12572, 1543, 50465], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 117, "seek": 32448, "start": 326.48, "end": 328.48, "text": "但把它轉成Q4量化板以後,", "tokens": [50465, 8395, 42061, 32739, 11336, 48, 19, 26748, 23756, 43664, 3588, 5661, 11, 50565], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 118, "seek": 32448, "start": 328.48, "end": 330.48, "text": "加載就沒問題了。", "tokens": [50565, 9990, 12290, 231, 3111, 4913, 17197, 2289, 1543, 50665], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 119, "seek": 32448, "start": 330.48, "end": 332.48, "text": "但是大參數模型轉換比較慢,", "tokens": [50665, 11189, 3582, 47375, 30622, 41908, 39823, 32739, 36338, 25174, 20645, 11, 50765], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 120, "seek": 32448, "start": 332.48, "end": 334.48, "text": "像我這次用了20多分鐘。", "tokens": [50765, 12760, 1654, 2664, 9487, 9254, 2289, 2009, 6392, 32611, 1543, 50865], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 121, "seek": 32448, "start": 334.48, "end": 336.48, "text": "當有GGUF時,", "tokens": [50865, 13118, 2412, 27561, 52, 37, 6611, 11, 50965], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 122, "seek": 32448, "start": 336.48, "end": 338.48, "text": "點這裡加載模型,", "tokens": [50965, 8216, 18771, 9990, 12290, 231, 41908, 39823, 11, 51065], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 123, "seek": 32448, "start": 338.48, "end": 340.48, "text": "後台服務會優先加載量化板。", "tokens": [51065, 5661, 15433, 27408, 39517, 6236, 12827, 103, 10108, 9990, 12290, 231, 26748, 23756, 43664, 1543, 51165], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 124, "seek": 32448, "start": 340.48, "end": 341.48, "text": "需要提醒一點,", "tokens": [51165, 35748, 20949, 40796, 19202, 11, 51215], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 125, "seek": 32448, "start": 341.48, "end": 345.48, "text": "目前AudioCPP並不支持其他框架轉換的GGUF,", "tokens": [51215, 39004, 15591, 1004, 34, 17755, 35934, 1960, 35488, 9572, 5000, 21416, 228, 7360, 114, 32739, 36338, 1546, 27561, 52, 37, 11, 51415], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 126, "seek": 32448, "start": 345.48, "end": 347.48, "text": "並且也只有部分模型現在可以轉換。", "tokens": [51415, 35934, 20334, 6404, 35244, 32174, 41908, 39823, 12648, 6723, 32739, 36338, 1543, 51515], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 127, "seek": 32448, "start": 347.48, "end": 349.48, "text": "在列表裡選模型時,", "tokens": [51515, 3581, 43338, 17571, 11066, 21625, 41908, 39823, 6611, 11, 51615], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 128, "seek": 32448, "start": 349.48, "end": 351.48, "text": "能不能轉換下面都有提示。", "tokens": [51615, 8225, 28590, 32739, 36338, 47150, 48121, 20949, 25696, 1543, 51715], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 129, "seek": 32448, "start": 351.48, "end": 353.48, "text": "轉好的量化板不想用也可以直接轉換。", "tokens": [51715, 32739, 20715, 26748, 23756, 43664, 1960, 7093, 9254, 47935, 43297, 32739, 36338, 1543, 51815], "temperature": 0, "avg_logprob": -0.0809206462038156, "compression_ratio": 1.336231884057971, "no_speech_prob": 1.792479706330674e-11}, {"id": 130, "seek": 35348, "start": 353.48, "end": 355.48, "text": "就能恢復原始權重。", "tokens": [50365, 3111, 8225, 11845, 95, 2172, 102, 19683, 15476, 40212, 12624, 1543, 50465], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 131, "seek": 35348, "start": 355.48, "end": 359.48, "text": "目前AudioCPP已經完整支持30個語音模型。", "tokens": [50465, 39004, 15591, 1004, 34, 17755, 25338, 14128, 27662, 35488, 3446, 3338, 31348, 18034, 41908, 39823, 1543, 50665], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 132, "seek": 35348, "start": 359.48, "end": 361.48, "text": "所有模型的能力特點,", "tokens": [50665, 39300, 41908, 39823, 1546, 8225, 13486, 17682, 8216, 11, 50765], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 133, "seek": 35348, "start": 361.48, "end": 363.48, "text": "我都整理成了配套筆記。", "tokens": [50765, 1654, 7182, 27662, 13876, 11336, 2289, 38846, 1881, 245, 7973, 228, 16958, 1543, 50865], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 134, "seek": 35348, "start": 363.48, "end": 365.48, "text": "可以通過這張表格來數查。", "tokens": [50865, 6723, 19550, 8816, 2664, 18889, 17571, 30921, 3763, 30622, 42623, 1543, 50965], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 135, "seek": 35348, "start": 365.48, "end": 367.48, "text": "鏈接我放在評論區置頂。", "tokens": [50965, 46241, 230, 14468, 1654, 12744, 3581, 39115, 29340, 44526, 34719, 8313, 224, 1543, 51065], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 136, "seek": 35348, "start": 367.48, "end": 369.48, "text": "最後說下整合包怎麼升級,", "tokens": [51065, 20578, 4622, 4438, 27662, 14245, 23305, 11614, 41670, 35315, 11, 51165], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 137, "seek": 35348, "start": 369.48, "end": 370.48, "text": "以及模型下載。", "tokens": [51165, 40282, 41908, 39823, 4438, 12290, 231, 1543, 51215], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 138, "seek": 35348, "start": 370.48, "end": 373.48, "text": "第一個鏈接是新的完整版整合包。", "tokens": [51215, 40760, 46241, 230, 14468, 1541, 12560, 1546, 14128, 27662, 42096, 27662, 14245, 23305, 1543, 51365], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 139, "seek": 35348, "start": 373.48, "end": 374.48, "text": "如果你之前沒安裝過,", "tokens": [51365, 45669, 32442, 4913, 16206, 44453, 8816, 11, 51415], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 140, "seek": 35348, "start": 374.48, "end": 375.48, "text": "就下這個。", "tokens": [51415, 3111, 4438, 6287, 1543, 51465], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 141, "seek": 35348, "start": 375.48, "end": 376.48, "text": "上期安裝過的,", "tokens": [51465, 5708, 16786, 16206, 44453, 8816, 1546, 11, 51515], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 142, "seek": 35348, "start": 376.48, "end": 378.48, "text": "點第二個鏈接下載升級包。", "tokens": [51515, 8216, 49085, 46241, 230, 14468, 4438, 12290, 231, 41670, 35315, 23305, 1543, 51615], "temperature": 0, "avg_logprob": -0.0895585013024601, "compression_ratio": 1.3003194888178913, "no_speech_prob": 1.730205215211278e-11}, {"id": 143, "seek": 37848, "start": 378.48, "end": 379.48, "text": "解壓以後,", "tokens": [50365, 17278, 45964, 3588, 5661, 11, 50415], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 144, "seek": 37848, "start": 379.48, "end": 380.48, "text": "複製裡面的所有文件,", "tokens": [50415, 164, 97, 229, 8083, 121, 32399, 1546, 39300, 17174, 20485, 11, 50465], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 145, "seek": 37848, "start": 380.48, "end": 382.48, "text": "覆蓋到原來整合包目錄下。", "tokens": [50465, 2888, 228, 39198, 233, 4511, 19683, 3763, 27662, 14245, 23305, 11386, 9567, 226, 4438, 1543, 50565], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 146, "seek": 37848, "start": 382.48, "end": 384.48, "text": "然後全部替換,", "tokens": [50565, 10213, 38714, 9531, 123, 36338, 11, 50665], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 147, "seek": 37848, "start": 384.48, "end": 385.48, "text": "就升級好了。", "tokens": [50665, 3111, 41670, 35315, 12621, 1543, 50715], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 148, "seek": 37848, "start": 385.48, "end": 386.48, "text": "以後更新,", "tokens": [50715, 3588, 5661, 19002, 12560, 11, 50765], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 149, "seek": 37848, "start": 386.48, "end": 387.48, "text": "可以直接點update。", "tokens": [50765, 6723, 43297, 8216, 1010, 17393, 1543, 50815], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 150, "seek": 37848, "start": 387.48, "end": 388.48, "text": "程序會自動檢測,", "tokens": [50815, 29649, 6346, 237, 6236, 9722, 12572, 40741, 95, 9592, 105, 11, 50865], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 151, "seek": 37848, "start": 388.48, "end": 390.48, "text": "有新版本會自動下載。", "tokens": [50865, 2412, 12560, 42096, 8802, 6236, 9722, 12572, 4438, 12290, 231, 1543, 50965], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 152, "seek": 37848, "start": 390.48, "end": 392.48, "text": "雙擊Run WebUI打開界面。", "tokens": [50965, 6306, 247, 37813, 42639, 9573, 46324, 12467, 8949, 21185, 8833, 1543, 51065], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 153, "seek": 37848, "start": 392.48, "end": 394.48, "text": "這個啟動窗口第一次要等幾十秒。", "tokens": [51065, 6287, 3284, 253, 12572, 30771, 245, 18144, 18049, 9487, 4275, 10187, 23575, 20145, 20956, 1543, 51165], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 154, "seek": 37848, "start": 394.48, "end": 396.48, "text": "程序加載完會自動打開網頁。", "tokens": [51165, 29649, 6346, 237, 9990, 12290, 231, 14128, 6236, 9722, 12572, 12467, 8949, 26352, 8313, 223, 1543, 51265], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 155, "seek": 37848, "start": 396.48, "end": 398.48, "text": "再說下模型下載。", "tokens": [51265, 8623, 4622, 4438, 41908, 39823, 4438, 12290, 231, 1543, 51365], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 156, "seek": 37848, "start": 398.48, "end": 400.48, "text": "目前AudioCPP的模型", "tokens": [51365, 39004, 15591, 1004, 34, 17755, 1546, 41908, 39823, 51465], "temperature": 0, "avg_logprob": -0.11889293717174995, "compression_ratio": 1.2173913043478262, "no_speech_prob": 1.8702473597587144e-11}, {"id": 157, "seek": 40048, "start": 400.48, "end": 402.48, "text": "都要從Hugging Face上下載。", "tokens": [50365, 7182, 4275, 21068, 39, 697, 3249, 4047, 5708, 4438, 12290, 231, 1543, 50465], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 158, "seek": 40048, "start": 402.48, "end": 404.48, "text": "並且不支持手動放模型。", "tokens": [50465, 35934, 20334, 1960, 35488, 11389, 12572, 12744, 41908, 39823, 1543, 50565], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 159, "seek": 40048, "start": 404.48, "end": 406.48, "text": "國內網絡如果下載有問題,", "tokens": [50565, 8053, 28472, 26352, 6948, 94, 13119, 4438, 12290, 231, 2412, 17197, 11, 50665], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 160, "seek": 40048, "start": 406.48, "end": 408.48, "text": "可以試一下這裡填上代理地址。", "tokens": [50665, 6723, 22099, 8861, 18771, 13331, 104, 5708, 19105, 13876, 10928, 14872, 222, 1543, 50765], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 161, "seek": 40048, "start": 408.48, "end": 409.48, "text": "另外有個別模型,", "tokens": [50765, 26202, 2412, 3338, 16158, 41908, 39823, 11, 50815], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 162, "seek": 40048, "start": 409.48, "end": 411.48, "text": "像Stable Audio這個系列,", "tokens": [50815, 12760, 4520, 712, 25706, 6287, 25368, 43338, 11, 50915], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 163, "seek": 40048, "start": 411.48, "end": 413.48, "text": "需要在Hugging Face上申請。", "tokens": [50915, 35748, 3581, 39, 697, 3249, 4047, 5708, 3526, 111, 16302, 1543, 51015], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 164, "seek": 40048, "start": 413.48, "end": 415.48, "text": "具體操作步驟是這樣。", "tokens": [51015, 39806, 23987, 13167, 235, 11914, 31429, 24023, 253, 49603, 1543, 51115], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 165, "seek": 40048, "start": 415.48, "end": 417.48, "text": "先註册一個免費的Hugging Face帳號,", "tokens": [51115, 10108, 5396, 119, 5676, 234, 8990, 2347, 235, 50071, 1546, 39, 697, 3249, 4047, 4845, 111, 18616, 11, 51215], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 166, "seek": 40048, "start": 417.48, "end": 419.48, "text": "然後到這個地址填寫申請表單。", "tokens": [51215, 10213, 4511, 6287, 10928, 14872, 222, 13331, 104, 4510, 104, 3526, 111, 16302, 17571, 24227, 1543, 51315], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 167, "seek": 40048, "start": 419.48, "end": 421.48, "text": "申請通過後,", "tokens": [51315, 3526, 111, 16302, 19550, 8816, 5661, 11, 51415], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 168, "seek": 40048, "start": 421.48, "end": 423.48, "text": "在典禮的Hugging Face頭像,", "tokens": [51415, 3581, 2347, 116, 18614, 106, 1546, 39, 697, 3249, 4047, 19921, 12760, 11, 51515], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 169, "seek": 40048, "start": 423.48, "end": 424.48, "text": "菜單底部訪問Token。", "tokens": [51515, 33025, 24227, 22816, 13470, 5396, 103, 11361, 51, 8406, 1543, 51565], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 170, "seek": 40048, "start": 424.48, "end": 426.48, "text": "創建一個新Token,", "tokens": [51565, 5935, 113, 34157, 8990, 12560, 51, 8406, 11, 51665], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 171, "seek": 40048, "start": 426.48, "end": 427.48, "text": "選紙讀。", "tokens": [51665, 21625, 7732, 247, 7422, 222, 1543, 51715], "temperature": 0, "avg_logprob": -0.11144632131942196, "compression_ratio": 1.3353293413173652, "no_speech_prob": 1.7309606872850658e-11}, {"id": 172, "seek": 42748, "start": 427.48, "end": 429.48, "text": "複製創建的Token,", "tokens": [50365, 164, 97, 229, 8083, 121, 5935, 113, 34157, 1546, 51, 8406, 11, 50465], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 173, "seek": 42748, "start": 429.48, "end": 431.48, "text": "粘貼到WebUI的Token框裡,", "tokens": [50465, 18393, 246, 11561, 120, 4511, 4360, 65, 46324, 1546, 51, 8406, 21416, 228, 11066, 11, 50565], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 174, "seek": 42748, "start": 431.48, "end": 433.48, "text": "就可以下載模型了。", "tokens": [50565, 36781, 4438, 12290, 231, 41908, 39823, 2289, 1543, 50665], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 175, "seek": 42748, "start": 433.48, "end": 435.48, "text": "模型下載機制,", "tokens": [50665, 41908, 39823, 4438, 12290, 231, 17543, 25491, 11, 50765], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 176, "seek": 42748, "start": 435.48, "end": 436.48, "text": "是先在Modus目錄下,", "tokens": [50765, 1541, 10108, 3581, 44, 32419, 11386, 9567, 226, 4438, 11, 50815], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 177, "seek": 42748, "start": 436.48, "end": 438.48, "text": "創建一個臨時下載目錄,", "tokens": [50815, 5935, 113, 34157, 8990, 6784, 101, 6611, 4438, 12290, 231, 11386, 9567, 226, 11, 50915], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 178, "seek": 42748, "start": 438.48, "end": 439.48, "text": "就是這個文件夾。", "tokens": [50915, 5620, 6287, 17174, 20485, 1787, 122, 1543, 50965], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 179, "seek": 42748, "start": 439.48, "end": 441.48, "text": "既然能看到正在下載的模型。", "tokens": [50965, 4479, 95, 5823, 8225, 18032, 15789, 3581, 4438, 12290, 231, 1546, 41908, 39823, 1543, 51065], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 180, "seek": 42748, "start": 441.48, "end": 443.48, "text": "全部下載完成後,", "tokens": [51065, 38714, 4438, 12290, 231, 41509, 5661, 11, 51165], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 181, "seek": 42748, "start": 443.48, "end": 446.48, "text": "程序會自動創建重命名模型文件夾。", "tokens": [51165, 29649, 6346, 237, 6236, 9722, 12572, 5935, 113, 34157, 12624, 25236, 15940, 41908, 39823, 17174, 20485, 1787, 122, 1543, 51315], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 182, "seek": 42748, "start": 446.48, "end": 448.48, "text": "下載過程中中斷也沒關係。", "tokens": [51315, 4438, 12290, 231, 8816, 29649, 5975, 5975, 4307, 115, 6404, 37816, 1543, 51415], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 183, "seek": 42748, "start": 448.48, "end": 450.48, "text": "繼續點下載按鈕,", "tokens": [51415, 38459, 8216, 4438, 12290, 231, 26613, 41597, 243, 11, 51515], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 184, "seek": 42748, "start": 450.48, "end": 451.48, "text": "會斷點續傳。", "tokens": [51515, 6236, 4307, 115, 8216, 29583, 40852, 1543, 51565], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 185, "seek": 42748, "start": 451.48, "end": 454.48, "text": "下載完成後點刷新,", "tokens": [51565, 4438, 12290, 231, 41509, 5661, 8216, 2437, 115, 12560, 11, 51715], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 186, "seek": 42748, "start": 454.48, "end": 455.48, "text": "就可以下載模型了。", "tokens": [51715, 36781, 4438, 12290, 231, 41908, 39823, 2289, 1543, 51765], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 187, "seek": 42748, "start": 455.48, "end": 457.48, "text": "程序也會自检模型文件", "tokens": [51765, 29649, 6346, 237, 6404, 6236, 9722, 16407, 222, 41908, 39823, 17174, 20485, 51865], "temperature": 0, "avg_logprob": -0.11538601386837843, "compression_ratio": 1.4932432432432432, "no_speech_prob": 1.7857998066417302e-11}, {"id": 188, "seek": 45748, "start": 457.48, "end": 458.48, "text": "的完整性。", "tokens": [50365, 1546, 14128, 27662, 21686, 1543, 50415], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 189, "seek": 45748, "start": 458.48, "end": 460.48, "text": "缺文件會提示模型不完整。", "tokens": [50415, 38109, 118, 17174, 20485, 6236, 20949, 25696, 41908, 39823, 1960, 14128, 27662, 1543, 50515], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 190, "seek": 45748, "start": 460.48, "end": 462.48, "text": "新商手的朋友,", "tokens": [50515, 12560, 45581, 11389, 1546, 19828, 11, 50615], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 191, "seek": 45748, "start": 462.48, "end": 464.48, "text": "想了解介面基礎操作,", "tokens": [50615, 7093, 2289, 17278, 30312, 8833, 26008, 14971, 236, 13167, 235, 11914, 11, 50715], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 192, "seek": 45748, "start": 464.48, "end": 465.48, "text": "以及其他常用模型,", "tokens": [50715, 40282, 9572, 5000, 11279, 9254, 41908, 39823, 11, 50765], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 193, "seek": 45748, "start": 465.48, "end": 468.48, "text": "可以看語音AI系統化的那期視頻。", "tokens": [50765, 6723, 4200, 31348, 18034, 48698, 25368, 33725, 23756, 1546, 4184, 16786, 27333, 29940, 1543, 50915], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 194, "seek": 45748, "start": 468.48, "end": 470.48, "text": "本期視頻的配套筆記,", "tokens": [50915, 8802, 16786, 27333, 29940, 1546, 38846, 1881, 245, 7973, 228, 16958, 11, 51015], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 195, "seek": 45748, "start": 470.48, "end": 471.48, "text": "鏈接我放在評論區置頂。", "tokens": [51015, 46241, 230, 14468, 1654, 12744, 3581, 39115, 29340, 44526, 34719, 8313, 224, 1543, 51065], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 196, "seek": 45748, "start": 471.48, "end": 473.48, "text": "整個包使用中的問題,", "tokens": [51065, 27662, 3338, 23305, 22982, 9254, 5975, 1546, 17197, 11, 51165], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 197, "seek": 45748, "start": 473.48, "end": 474.48, "text": "或者好的點子想法,", "tokens": [51165, 31148, 20715, 8216, 7626, 7093, 11148, 11, 51215], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 198, "seek": 45748, "start": 474.48, "end": 476.48, "text": "都可以在評論區留言。", "tokens": [51215, 7182, 6723, 3581, 39115, 29340, 44526, 43681, 1543, 51315], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 199, "seek": 45748, "start": 476.48, "end": 478.48, "text": "我會定期收集整理,更新版本。", "tokens": [51315, 1654, 6236, 12088, 16786, 18681, 26020, 27662, 13876, 11, 19002, 12560, 42096, 8802, 1543, 51415], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 200, "seek": 45748, "start": 478.48, "end": 480.48, "text": "這裡是AI探索與發現,", "tokens": [51415, 18771, 1541, 48698, 6900, 95, 7732, 95, 34127, 38927, 11, 51515], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 201, "seek": 45748, "start": 480.48, "end": 481.48, "text": "感謝觀看,下期見。", "tokens": [51515, 40143, 25179, 4200, 11, 4438, 16786, 10470, 1543, 51565], "temperature": 0, "avg_logprob": -0.10592512857346308, "compression_ratio": 1.2793650793650793, "no_speech_prob": 1.7852757466796376e-11}, {"id": 202, "seek": 48748, "start": 487.48, "end": 489.48, "text": "我們在評論區,", "tokens": [50365, 5884, 3581, 39115, 29340, 44526, 11, 50465], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 203, "seek": 48748, "start": 489.48, "end": 490.48, "text": "我們在評論區,", "tokens": [50465, 5884, 3581, 39115, 29340, 44526, 11, 50515], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 204, "seek": 48748, "start": 490.48, "end": 491.48, "text": "我們在評論區,", "tokens": [50515, 5884, 3581, 39115, 29340, 44526, 11, 50565], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 205, "seek": 48748, "start": 491.48, "end": 492.48, "text": "我們在評論區,", "tokens": [50565, 5884, 3581, 39115, 29340, 44526, 11, 50615], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 206, "seek": 48748, "start": 492.48, "end": 493.48, "text": "我們在評論區,", "tokens": [50615, 5884, 3581, 39115, 29340, 44526, 11, 50665], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 207, "seek": 48748, "start": 493.48, "end": 494.48, "text": "我們在評論區,", "tokens": [50665, 5884, 3581, 39115, 29340, 44526, 11, 50715], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 208, "seek": 48748, "start": 494.48, "end": 495.48, "text": "我們在評論區,", "tokens": [50715, 5884, 3581, 39115, 29340, 44526, 11, 50765], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 209, "seek": 48748, "start": 495.48, "end": 496.48, "text": "我們在評論區,", "tokens": [50765, 5884, 3581, 39115, 29340, 44526, 11, 50815], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 210, "seek": 48748, "start": 496.48, "end": 497.48, "text": "我們在評論區,", "tokens": [50815, 5884, 3581, 39115, 29340, 44526, 11, 50865], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 211, "seek": 48748, "start": 497.48, "end": 498.48, "text": "我們在評論區,", "tokens": [50865, 5884, 3581, 39115, 29340, 44526, 11, 50915], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 212, "seek": 48748, "start": 498.48, "end": 499.48, "text": "我們在評論區,", "tokens": [50915, 5884, 3581, 39115, 29340, 44526, 11, 50965], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 213, "seek": 48748, "start": 499.48, "end": 500.48, "text": "我們在評論區,", "tokens": [50965, 5884, 3581, 39115, 29340, 44526, 11, 51015], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 214, "seek": 48748, "start": 500.48, "end": 501.48, "text": "我們在評論區,", "tokens": [51015, 5884, 3581, 39115, 29340, 44526, 11, 51065], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 215, "seek": 48748, "start": 501.48, "end": 502.48, "text": "我們在評論區,", "tokens": [51065, 5884, 3581, 39115, 29340, 44526, 11, 51115], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 216, "seek": 48748, "start": 502.48, "end": 503.48, "text": "我們在評論區,", "tokens": [51115, 5884, 3581, 39115, 29340, 44526, 11, 51165], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 217, "seek": 48748, "start": 503.48, "end": 504.48, "text": "我們在評論區,", "tokens": [51165, 5884, 3581, 39115, 29340, 44526, 11, 51215], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}, {"id": 218, "seek": 48748, "start": 504.48, "end": 505.48, "text": "我們在評論區,", "tokens": [51215, 5884, 3581, 39115, 29340, 44526, 11, 51265], "temperature": 0, "avg_logprob": -0.38256581624348956, "compression_ratio": 9.5, "no_speech_prob": 4.3028741680828375e-10}], "language": "zh"}