{ "text": "騰訊這回開源的不是聊天助手,是個知識庫,叫WeKnora。\n你電腦裡躺著幾百份文檔,硬盤越買越大,找起東西還是靠一頁頁翻,官方倉庫上就擺著這套東西。\n點開就是滿屏的說明文檔和安裝步驟。你是不是也想過,攢了這麼多資料,到底有沒有辦法一次管起來?\n官方中文站首頁一行字,騰訊開源,MIT,24000多星,全寫在一起。\n許可證是最寬鬆的那類,商用也隨便。說明文檔裡寫著,微信那邊對話開放平台就架在這套底子上面。\n騰訊把撐著微信那邊問答服務的底子白送了出來。可你自己那堆文檔,真接得上嗎?\n我自己也是被一堆文檔折磨過的人,所以把官方倉庫裡的說明文檔從頭翻到尾。\n這十分鐘,先說Week Nora是什麼,能替你幹哪些事。\n再說,她和直接把文檔丟給聊天礦的做法差在哪?\n最後說,官方自己寫明的短板,聽完再決定要不要折的。\n先說它是什麼,它專門啃印文檔。\n官方給的中文定義裡寫的是,結構複雜,內容易夠的文檔場景。\n翻成人話,就是板式亂七八糟,長短不一的那種文檔堆。\n結構越亂,越是她要啃的。\n官網首頁最大內行字就一句話,把散落文檔變成會思考的知識資產。\n這話不是我總結的,是她自己掛在那兒的。\n網上一行還寫著,騰訊開源企業級知識框架,\nRAG問答,Agent推理,Wiki圖譜,三檔合成一體。\n翻譯成人話,就是隨時問答,自己分多步找,自動養成知識頁面。\n隨時問答,這一檔,文檔放進去,直接問答案。\n客服的常見問題,這麼用。\n課程資料,也這麼用。\n平時查一條,翻半天。\n有了她,張嘴就問,還有個會自己分多步找的助手。\n問題複雜就自己拆步驟,還能調工具,連網搜一把也行。\n你只管把問題說清楚。\n中間那幾步,她自己跑,最厲害的是知識頁面那檔。\n散文檔自動熬成一套互相跳轉的頁面,能手改,有版本歷史,改壞了一鍵回滾。\n等於說她替你把一堆散文件,養成了一套能自己長大的資料站。\n她還長著記性,記得你是誰,記得你常問什麼,記之前還要你點頭確認,下次直接接著上回講,不用從頭介紹自己。\n官網上還專門掛了一句,答案自動標註來源,規避AI幻覺。\n就是不讓她一本正經瞎編。\n具體是哪份文檔的哪一段,一點就跳過去,敢不敢信她,全看這一條。\n還有一句,讓每份文檔可被檢索,可被推理,可自動整編。\n落到你身上就是找得到,問得動,不用你動手整理。\n知識分三類收,常見問題,文檔,知識頁,收進來自動打標籤,找的時候按標籤篩。\n公司那幾百份沒人看得完的文檔,扔進去就是個隨時能問的庫。\n考場把一整套資料丟進去,考前只管問,三千多次被人搬回去自己改,不是只有點心為觀的。\n聽著,像直接把文檔丟給聊天框,可兩邊根本不是一回事。\n差就差在誰替你管這堆東西。\n你平時把文檔丟給聊天框?餵一次答一次,那次答完他就忘了。\n下次新開一局,之前餵的全白費。\nWeek Nora這邊十幾種格式都能吃。\nPDF,文檔,表格,幻燈片,網頁存檔,甚至圖片和思維導圖。\n你手裡是什麼樣,扔進去就是什麼樣,不用先轉格式。\n整個文件夾拖進去,原來什麼目錄結構進來還是什麼樣,攢了多年的資料堆,原樣搬進來就行。\n找出來的那一段,要是不對,你可以直接改,改完它自己重新建一遍索引。\n酷是越用越準的。\n問答能掛到自己網站上,官方管這個叫域名白名單,就是只有你認的網站能接,還能限流。\n開店的,把問答掛到店裡,重複的問題不用一遍遍答,也能接進群裡。\n企業微信,飛書,釘釘,微信,都行,在群裡問一句,答案直接回在群裡。\n帶新人的群最合適。\n同樣的問題不用答三遍。\n刷到有用的網頁,瀏覽器擴展,選中就存進庫,官方管這個叫第二大腦。\n手機上還有小程序,接著問,晚上躺床上,接著查。\n多提一句向量這個詞,你可以理解成給每段文字編的號,機器靠這個號按意思找東西。\n號存在專門的數據庫裡,就是它放資料的那個倉庫,七八家可選。\n文檔和模型都能留在自己機器上,這是你自己的庫,不用每次重新喂一遍。\n官網演示裡,拿競品分析報告當例子,一句話讓它整理。\n還展示從三個知識庫,照回十幾條相關片段,再調數據看板,讀本季度銷售指標。\n這段是官網上的演示,不是我實測出來的。擺出來是給你看它想幹什麼活。\n東西是好東西,上手到底要折騰到什麼程度。\n先說最實在的上手路,把它下下來。\n改一行配置,敲兩句,啟動命令,瀏覽器打開,就是個能用的網站。\n前後四行,多一行都沒有。\n前置條件只有兩樣。\n一種把軟件連環境一起打包好,敲兩句就能啟動的工具。\n再加一個把代碼拉下來的小工具。\n嫌敲命令麻煩。\n官網還有個騰訊雲一鍵部署的按鈕。\n一鍵這兩個字,是人家自己寫的。\n但這條走的是騰訊雲上的服務,不是跑在你自己電腦上。\n要不要錢,怎麼算。\n你得上那頁自己看。\n前置條件那一欄裡,一個顯存數字都沒提。\n因為模型不在它身上跑。\n你想全離線,顯卡的賬得算在你選的那個本地模型頭上。\n走本地模型工具,也就是一個能在自己電腦上跑開源模型的工具。\n文檔不出自己電腦,模型也在自己電腦上跑。\n這條路官方顯明了支持,線上模型支持20來家。\n國產幾家也在裡面。\n不過默認接的是線上模型,全離線,要自己換成本地那套。\n別裝完就以為已經斷網了。\n還有一條壓根不用自己裝的路,\n騰訊把這套東西做成了微信的一個對話開放平台。\n傳知識就能用,不用碰命令行,也就是不用敲那幾行字。\n連裝都不想裝的人走這條就行。\n官方另外準備了一個輕量版,有單獨的安裝配方。\n能省多少?跟完整版差在哪?說明文檔裡沒寫,我也不替它吹。\n等官方把話說全再講。\n自己攢了多年的電子書堆,整個目錄拖進去。\n出門路上想查就查。\n好處說完了,那代價呢?\n白送的是軟件,運維是你自己的。\n官方原話寫在安全提示裡,裝在內部網自己用,別直接掛到公網,掛出去會漏東西。\n這是他自己寫的,不是我嚇唬你。\n登錄驗證是後來才補上的,最早的版本連登錄都沒有,這條也寫在同一段安全提示裡。\n門鎖是後裝的,早下手的人裸奔過一陣。\n讓AI在你機器上動手腳的地方,官方一直在收緊,默認關著,手動才能開。\n剛到0.8版,能動手腳的路又砍掉一條,這點我支持官方。\n角色權限分四檔,還有審計日誌,多工作區,邀請鏈,動了哪份文檔,誰動的,日誌裡都記著。\n一個人的資料庫,被迫先讀懂一套管幾百人的規矩。\n裝起來不是一個小軟件,是一套服務。存編號的數據庫要配,放文件的,排任務的也得配上,拼起來才能轉。\n愛折騰的人覺得是自由,怕麻煩的人看到就頭大。\n防火牆和訪問控制自己配,補丁自己打,管這套東西的人就是你自己。\n半夜出問題,搖不到騰訊的人,只能自己爬起來看。\n三百多條問題還開著,將近三百個沒合的改動請求。\n活躍是活躍,坑也不少。\n所以我的判斷是,文檔多到,翻不動的人值得折騰。\n手裡只有幾十份文檔,真沒必要折騰。\n先從微信那邊免部署那條路,順手再自己裝。\n買不買賬,取決於你手裡文檔有多少。\n你們那幾百份文檔不用再一份份翻了。\n騰訊開源了,一套能把它們變成隨時能問的知識庫。\n文檔還能不出公司,這類能自己裝起來的國產開源工具。\n我會一個個替你試過再講。\n包括那些其實跑不起來的。\n最後問你一句,公司那堆文檔你是想自己裝一套慢慢問,\n還是直接傳知識用現成的?\n裝過的講講卡在哪一步?\n留言區的評論區報個一或者二,你選哪個?", "language": "zh", "duration": 554.581375, "segments": [ { "id": 0, "seek": 0, "start": 0.0, "end": 5.2, "text": "騰訊這回開源的不是聊天助手,是個知識庫,叫WeKnora。", "tokens": [ 50365, 38168, 108, 5396, 232, 2664, 8350, 8949, 47402, 1546, 7296, 40096, 6135, 13557, 11389, 11, 1541, 3338, 6498, 43143, 6346, 104, 11, 19855, 4360, 916, 45741, 1543, 50625 ], "temperature": 0, "avg_logprob": -0.15569175779819489, "compression_ratio": 1.158450704225352, "no_speech_prob": 9.934789969856528e-12, "source": "whisper", "text_original": "騰訊這回開源的不是聊天主手,是個知識庫,叫Week Nora。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 1, "seek": 0, "start": 5.2, "end": 14.6, "text": "你電腦裡躺著幾百份文檔,硬盤越買越大,找起東西還是靠一頁頁翻,官方倉庫上就擺著這套東西。", "tokens": [ 50625, 2166, 20545, 21184, 99, 11066, 17514, 118, 19382, 23575, 31906, 36266, 17174, 40741, 242, 11, 38114, 105, 5419, 97, 25761, 31001, 25761, 3582, 11, 25085, 9147, 26978, 25583, 5363, 254, 2257, 8313, 223, 8313, 223, 42716, 11, 31929, 9249, 2122, 231, 6346, 104, 5708, 3111, 13167, 118, 19382, 2664, 1881, 245, 26978, 1543, 51095 ], "temperature": 0, "avg_logprob": -0.15569175779819489, "compression_ratio": 1.158450704225352, "no_speech_prob": 9.934789969856528e-12, "source": "whisper" }, { "id": 2, "seek": 0, "start": 14.6, "end": 23.8, "text": "點開就是滿屏的說明文檔和安裝步驟。你是不是也想過,攢了這麼多資料,到底有沒有辦法一次管起來?", "tokens": [ 51095, 8216, 8949, 5620, 32197, 9636, 237, 1546, 4622, 11100, 17174, 40741, 242, 12565, 16206, 44453, 31429, 24023, 253, 1543, 2166, 23034, 6404, 7093, 8816, 11, 5917, 95, 2289, 21269, 6392, 35380, 33404, 11, 33883, 28350, 40072, 27505, 23131, 21670, 30, 51555 ], "temperature": 0, "avg_logprob": -0.15569175779819489, "compression_ratio": 1.158450704225352, "no_speech_prob": 9.934789969856528e-12, "source": "whisper" }, { "id": 3, "seek": 2380, "start": 23.8, "end": 30.400000000000002, "text": "官方中文站首頁一行字,騰訊開源,MIT,24000多星,全寫在一起。", "tokens": [ 50365, 31929, 9249, 5975, 17174, 34155, 25444, 8313, 223, 2257, 8082, 22381, 11, 38168, 108, 5396, 232, 8949, 47402, 11, 44, 3927, 11, 7911, 1360, 6392, 20682, 11, 11319, 4510, 104, 3581, 29567, 1543, 50695 ], "temperature": 0, "avg_logprob": -0.10968081523917314, "compression_ratio": 1.3052959501557633, "no_speech_prob": 1.1478921979612977e-11, "source": "whisper" }, { "id": 4, "seek": 2380, "start": 30.400000000000002, "end": 39.4, "text": "許可證是最寬鬆的那類,商用也隨便。說明文檔裡寫著,微信那邊對話開放平台就架在這套底子上面。", "tokens": [ 50695, 40914, 4429, 11823, 231, 1541, 8661, 4510, 105, 32506, 228, 1546, 4184, 40092, 11, 45581, 9254, 6404, 48890, 27364, 1543, 4622, 11100, 17174, 40741, 242, 11066, 4510, 104, 19382, 11, 39152, 17665, 4184, 14901, 2855, 11103, 8949, 12744, 16716, 15433, 3111, 7360, 114, 3581, 2664, 1881, 245, 22816, 7626, 49750, 1543, 51145 ], "temperature": 0, "avg_logprob": -0.10968081523917314, "compression_ratio": 1.3052959501557633, "no_speech_prob": 1.1478921979612977e-11, "source": "whisper" }, { "id": 5, "seek": 2380, "start": 39.4, "end": 46.8, "text": "騰訊把撐著微信那邊問答服務的底子白送了出來。可你自己那堆文檔,真接得上嗎?", "tokens": [ 51145, 38168, 108, 5396, 232, 16075, 20559, 238, 19382, 39152, 17665, 4184, 14901, 11361, 28223, 27408, 39517, 1546, 22816, 7626, 13558, 29309, 2289, 29741, 1543, 4429, 2166, 17645, 4184, 10726, 228, 17174, 40741, 242, 11, 6303, 14468, 5916, 5708, 7434, 30, 51515 ], "temperature": 0, "avg_logprob": -0.10968081523917314, "compression_ratio": 1.3052959501557633, "no_speech_prob": 1.1478921979612977e-11, "source": "whisper" }, { "id": 6, "seek": 2380, "start": 46.8, "end": 53.6, "text": "我自己也是被一堆文檔折磨過的人,所以把官方倉庫裡的說明文檔從頭翻到尾。", "tokens": [ 51515, 1654, 17645, 22021, 23238, 2257, 10726, 228, 17174, 40741, 242, 7235, 246, 163, 13033, 8816, 29979, 11, 7239, 16075, 31929, 9249, 2122, 231, 6346, 104, 11066, 1546, 4622, 11100, 17174, 40741, 242, 21068, 19921, 42716, 4511, 1530, 122, 1543, 51855 ], "temperature": 0, "avg_logprob": -0.10968081523917314, "compression_ratio": 1.3052959501557633, "no_speech_prob": 1.1478921979612977e-11, "source": "whisper" }, { "id": 7, "seek": 5360, "start": 53.6, "end": 58.6, "text": "這十分鐘,先說Week Nora是什麼,能替你幹哪些事。", "tokens": [ 50365, 2664, 20145, 32611, 11, 10108, 4622, 4360, 916, 45741, 35851, 11, 8225, 9531, 123, 2166, 29932, 17028, 13824, 6973, 1543, 50615 ], "temperature": 0, "avg_logprob": -0.16916980383531102, "compression_ratio": 1.2388535031847134, "no_speech_prob": 1.7437490687499668e-11, "source": "whisper" }, { "id": 8, "seek": 5360, "start": 58.6, "end": 62.6, "text": "再說,她和直接把文檔丟給聊天礦的做法差在哪?", "tokens": [ 50615, 8623, 4622, 11, 15555, 12565, 43297, 16075, 17174, 40741, 242, 940, 253, 17798, 40096, 6135, 14971, 99, 1546, 10907, 11148, 21679, 3581, 17028, 30, 50815 ], "temperature": 0, "avg_logprob": -0.16916980383531102, "compression_ratio": 1.2388535031847134, "no_speech_prob": 1.7437490687499668e-11, "source": "whisper" }, { "id": 9, "seek": 5360, "start": 62.6, "end": 67.6, "text": "最後說,官方自己寫明的短板,聽完再決定要不要折的。", "tokens": [ 50815, 20578, 4622, 11, 31929, 9249, 17645, 4510, 104, 11100, 1546, 5881, 255, 43664, 11, 21524, 14128, 8623, 33540, 12088, 4275, 11962, 7235, 246, 1546, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.16916980383531102, "compression_ratio": 1.2388535031847134, "no_speech_prob": 1.7437490687499668e-11, "source": "whisper" }, { "id": 10, "seek": 5360, "start": 67.6, "end": 70.6, "text": "先說它是什麼,它專門啃印文檔。", "tokens": [ 51065, 10108, 4622, 11284, 35851, 11, 11284, 46072, 37232, 3284, 225, 35825, 17174, 40741, 242, 1543, 51215 ], "temperature": 0, "avg_logprob": -0.16916980383531102, "compression_ratio": 1.2388535031847134, "no_speech_prob": 1.7437490687499668e-11, "source": "whisper" }, { "id": 11, "seek": 5360, "start": 70.6, "end": 76.6, "text": "官方給的中文定義裡寫的是,結構複雜,內容易夠的文檔場景。", "tokens": [ 51215, 31929, 9249, 17798, 1546, 5975, 17174, 12088, 37430, 11066, 4510, 104, 24620, 11, 17144, 43362, 164, 97, 229, 6306, 250, 11, 28472, 25750, 31962, 31649, 1546, 17174, 40741, 242, 21446, 50218, 1543, 51515 ], "temperature": 0, "avg_logprob": -0.16916980383531102, "compression_ratio": 1.2388535031847134, "no_speech_prob": 1.7437490687499668e-11, "source": "whisper" }, { "id": 12, "seek": 5360, "start": 76.6, "end": 80.6, "text": "翻成人話,就是板式亂七八糟,長短不一的那種文檔堆。", "tokens": [ 51515, 42716, 11336, 4035, 11103, 11, 5620, 43664, 27584, 1369, 224, 29920, 33453, 15239, 253, 11, 15353, 5881, 255, 1960, 2257, 1546, 4184, 16516, 17174, 40741, 242, 10726, 228, 1543, 51715 ], "temperature": 0, "avg_logprob": -0.16916980383531102, "compression_ratio": 1.2388535031847134, "no_speech_prob": 1.7437490687499668e-11, "source": "whisper" }, { "id": 13, "seek": 8060, "start": 80.6, "end": 83.6, "text": "結構越亂,越是她要啃的。", "tokens": [ 50365, 17144, 43362, 25761, 1369, 224, 11, 25761, 1541, 15555, 4275, 3284, 225, 1546, 1543, 50515 ], "temperature": 0, "avg_logprob": -0.13444722841863763, "compression_ratio": 1.0858895705521472, "no_speech_prob": 1.897317546128985e-11, "source": "whisper" }, { "id": 14, "seek": 8060, "start": 83.6, "end": 89.6, "text": "官網首頁最大內行字就一句話,把散落文檔變成會思考的知識資產。", "tokens": [ 50515, 31929, 26352, 25444, 8313, 223, 8661, 3582, 28472, 8082, 22381, 3111, 2257, 34592, 11103, 11, 16075, 7017, 96, 30848, 17174, 40741, 242, 26327, 11336, 6236, 8870, 26504, 1546, 6498, 43143, 35380, 33299, 1543, 50815 ], "temperature": 0, "avg_logprob": -0.13444722841863763, "compression_ratio": 1.0858895705521472, "no_speech_prob": 1.897317546128985e-11, "source": "whisper" }, { "id": 15, "seek": 8060, "start": 89.6, "end": 92.6, "text": "這話不是我總結的,是她自己掛在那兒的。", "tokens": [ 50815, 2664, 11103, 7296, 1654, 26575, 17144, 1546, 11, 1541, 15555, 17645, 6900, 249, 3581, 4184, 35125, 1546, 1543, 50965 ], "temperature": 0, "avg_logprob": -0.13444722841863763, "compression_ratio": 1.0858895705521472, "no_speech_prob": 1.897317546128985e-11, "source": "whisper" }, { "id": 16, "seek": 9260, "start": 92.6, "end": 96.6, "text": "網上一行還寫著,騰訊開源企業級知識框架,", "tokens": [ 50365, 26352, 5708, 2257, 8082, 7824, 4510, 104, 19382, 11, 38168, 108, 5396, 232, 8949, 47402, 7384, 223, 27119, 35315, 6498, 43143, 21416, 228, 7360, 114, 11, 50565 ], "temperature": 0, "avg_logprob": -0.11559813731425517, "compression_ratio": 1.2158273381294964, "no_speech_prob": 1.6131403504648922e-11, "source": "whisper" }, { "id": 17, "seek": 9260, "start": 96.6, "end": 100.6, "text": "RAG問答,Agent推理,Wiki圖譜,三檔合成一體。", "tokens": [ 50565, 3750, 34, 11361, 28223, 11, 32, 6930, 33597, 13876, 11, 54, 9850, 2523, 244, 11823, 250, 11, 10960, 40741, 242, 14245, 11336, 2257, 23987, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.11559813731425517, "compression_ratio": 1.2158273381294964, "no_speech_prob": 1.6131403504648922e-11, "source": "whisper", "text_original": "RAC問答,Agent推理,Wiki圖譜,三檔合成一體。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 18, "seek": 9260, "start": 100.6, "end": 106.6, "text": "翻譯成人話,就是隨時問答,自己分多步找,自動養成知識頁面。", "tokens": [ 50765, 42716, 11823, 107, 11336, 4035, 11103, 11, 5620, 48890, 6611, 11361, 28223, 11, 17645, 6627, 6392, 31429, 25085, 11, 9722, 12572, 22217, 232, 11336, 6498, 43143, 8313, 223, 8833, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.11559813731425517, "compression_ratio": 1.2158273381294964, "no_speech_prob": 1.6131403504648922e-11, "source": "whisper" }, { "id": 19, "seek": 9260, "start": 106.6, "end": 110.6, "text": "隨時問答,這一檔,文檔放進去,直接問答案。", "tokens": [ 51065, 48890, 6611, 11361, 28223, 11, 32260, 40741, 242, 11, 17174, 40741, 242, 12744, 18214, 6734, 11, 43297, 11361, 28223, 28899, 1543, 51265 ], "temperature": 0, "avg_logprob": -0.11559813731425517, "compression_ratio": 1.2158273381294964, "no_speech_prob": 1.6131403504648922e-11, "source": "whisper" }, { "id": 20, "seek": 9260, "start": 110.6, "end": 112.6, "text": "客服的常見問題,這麼用。", "tokens": [ 51265, 32316, 27408, 1546, 11279, 10470, 17197, 11, 21269, 9254, 1543, 51365 ], "temperature": 0, "avg_logprob": -0.11559813731425517, "compression_ratio": 1.2158273381294964, "no_speech_prob": 1.6131403504648922e-11, "source": "whisper" }, { "id": 21, "seek": 9260, "start": 112.6, "end": 113.6, "text": "課程資料,也這麼用。", "tokens": [ 51365, 3549, 110, 29649, 35380, 33404, 11, 6404, 21269, 9254, 1543, 51415 ], "temperature": 0, "avg_logprob": -0.11559813731425517, "compression_ratio": 1.2158273381294964, "no_speech_prob": 1.6131403504648922e-11, "source": "whisper" }, { "id": 22, "seek": 9260, "start": 113.6, "end": 115.6, "text": "平時查一條,翻半天。", "tokens": [ 51415, 16716, 6611, 42623, 2257, 31106, 11, 42716, 30018, 6135, 1543, 51515 ], "temperature": 0, "avg_logprob": -0.11559813731425517, "compression_ratio": 1.2158273381294964, "no_speech_prob": 1.6131403504648922e-11, "source": "whisper" }, { "id": 23, "seek": 11560, "start": 115.6, "end": 120.6, "text": "有了她,張嘴就問,還有個會自己分多步找的助手。", "tokens": [ 50365, 2412, 2289, 15555, 11, 18889, 10756, 112, 3111, 11361, 11, 15569, 3338, 6236, 17645, 6627, 6392, 31429, 25085, 1546, 37618, 11389, 1543, 50615 ], "temperature": 0, "avg_logprob": -0.07481825011117117, "compression_ratio": 1.264367816091954, "no_speech_prob": 1.8742811122574032e-11, "source": "whisper" }, { "id": 24, "seek": 11560, "start": 120.6, "end": 125.6, "text": "問題複雜就自己拆步驟,還能調工具,連網搜一把也行。", "tokens": [ 50615, 17197, 164, 97, 229, 6306, 250, 3111, 17645, 6852, 228, 31429, 24023, 253, 11, 7824, 8225, 32943, 23323, 39806, 11, 36252, 26352, 162, 14987, 2257, 16075, 6404, 8082, 1543, 50865 ], "temperature": 0, "avg_logprob": -0.07481825011117117, "compression_ratio": 1.264367816091954, "no_speech_prob": 1.8742811122574032e-11, "source": "whisper" }, { "id": 25, "seek": 11560, "start": 125.6, "end": 127.6, "text": "你只管把問題說清楚。", "tokens": [ 50865, 2166, 14003, 23131, 16075, 17197, 4622, 38433, 1543, 50965 ], "temperature": 0, "avg_logprob": -0.07481825011117117, "compression_ratio": 1.264367816091954, "no_speech_prob": 1.8742811122574032e-11, "source": "whisper" }, { "id": 26, "seek": 11560, "start": 127.6, "end": 131.6, "text": "中間那幾步,她自己跑,最厲害的是知識頁面那檔。", "tokens": [ 50965, 5975, 11016, 4184, 23575, 31429, 11, 15555, 17645, 32585, 11, 8661, 4510, 102, 27333, 24620, 6498, 43143, 8313, 223, 8833, 4184, 40741, 242, 1543, 51165 ], "temperature": 0, "avg_logprob": -0.07481825011117117, "compression_ratio": 1.264367816091954, "no_speech_prob": 1.8742811122574032e-11, "source": "whisper", "text_original": "中間那幾步,她自己跑,最審視的是知識頁面那檔。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 27, "seek": 11560, "start": 131.6, "end": 138.6, "text": "散文檔自動熬成一套互相跳轉的頁面,能手改,有版本歷史,改壞了一鍵回滾。", "tokens": [ 51165, 7017, 96, 17174, 40741, 242, 9722, 12572, 21805, 105, 11336, 2257, 1881, 245, 1369, 240, 15106, 29366, 32739, 1546, 8313, 223, 8833, 11, 8225, 11389, 34490, 11, 2412, 42096, 8802, 4287, 115, 45399, 11, 34490, 12829, 252, 2289, 2257, 41668, 113, 8350, 15868, 122, 1543, 51515 ], "temperature": 0, "avg_logprob": -0.07481825011117117, "compression_ratio": 1.264367816091954, "no_speech_prob": 1.8742811122574032e-11, "source": "whisper" }, { "id": 28, "seek": 13860, "start": 138.6, "end": 144.6, "text": "等於說她替你把一堆散文件,養成了一套能自己長大的資料站。", "tokens": [ 50365, 10187, 19488, 4622, 15555, 9531, 123, 2166, 16075, 2257, 10726, 228, 7017, 96, 17174, 20485, 11, 22217, 232, 11336, 2289, 2257, 1881, 245, 8225, 17645, 15353, 39156, 35380, 33404, 34155, 1543, 50665 ], "temperature": 0, "avg_logprob": -0.07952660548536084, "compression_ratio": 1.2953846153846154, "no_speech_prob": 1.8482862809698908e-11, "source": "whisper" }, { "id": 29, "seek": 13860, "start": 144.6, "end": 153.6, "text": "她還長著記性,記得你是誰,記得你常問什麼,記之前還要你點頭確認,下次直接接著上回講,不用從頭介紹自己。", "tokens": [ 50665, 15555, 7824, 15353, 19382, 16958, 21686, 11, 30872, 32526, 20757, 11, 30872, 2166, 11279, 11361, 7598, 11, 16958, 32442, 7824, 4275, 2166, 8216, 19921, 24293, 22041, 11, 35829, 43297, 14468, 19382, 5708, 8350, 11932, 11, 24384, 21068, 19921, 30312, 38193, 17645, 1543, 51115 ], "temperature": 0, "avg_logprob": -0.07952660548536084, "compression_ratio": 1.2953846153846154, "no_speech_prob": 1.8482862809698908e-11, "source": "whisper" }, { "id": 30, "seek": 13860, "start": 153.6, "end": 158.6, "text": "官網上還專門掛了一句,答案自動標註來源,規避AI幻覺。", "tokens": [ 51115, 31929, 26352, 5708, 7824, 46072, 37232, 6900, 249, 2289, 2257, 34592, 11, 28223, 28899, 9722, 12572, 46299, 5396, 119, 3763, 47402, 11, 2888, 237, 3330, 123, 48698, 3509, 119, 9463, 1543, 51365 ], "temperature": 0, "avg_logprob": -0.07952660548536084, "compression_ratio": 1.2953846153846154, "no_speech_prob": 1.8482862809698908e-11, "source": "whisper" }, { "id": 31, "seek": 13860, "start": 158.6, "end": 160.6, "text": "就是不讓她一本正經瞎編。", "tokens": [ 51365, 5620, 1960, 21195, 15555, 2257, 8802, 15789, 15145, 36733, 236, 12608, 101, 1543, 51465 ], "temperature": 0, "avg_logprob": -0.07952660548536084, "compression_ratio": 1.2953846153846154, "no_speech_prob": 1.8482862809698908e-11, "source": "whisper" }, { "id": 32, "seek": 13860, "start": 160.6, "end": 166.6, "text": "具體是哪份文檔的哪一段,一點就跳過去,敢不敢信她,全看這一條。", "tokens": [ 51465, 39806, 23987, 1541, 17028, 36266, 17174, 40741, 242, 1546, 17028, 2257, 28427, 11, 19202, 3111, 29366, 36910, 11, 40205, 1960, 40205, 17665, 15555, 11, 11319, 4200, 32260, 31106, 1543, 51765 ], "temperature": 0, "avg_logprob": -0.07952660548536084, "compression_ratio": 1.2953846153846154, "no_speech_prob": 1.8482862809698908e-11, "source": "whisper" }, { "id": 33, "seek": 16660, "start": 166.6, "end": 172.6, "text": "還有一句,讓每份文檔可被檢索,可被推理,可自動整編。", "tokens": [ 50365, 15569, 2257, 34592, 11, 21195, 23664, 36266, 17174, 40741, 242, 4429, 23238, 40741, 95, 7732, 95, 11, 4429, 23238, 33597, 13876, 11, 4429, 9722, 12572, 27662, 12608, 101, 1543, 50665 ], "temperature": 0, "avg_logprob": -0.06698141911233119, "compression_ratio": 1.3008130081300813, "no_speech_prob": 2.0075172427191923e-11, "source": "whisper" }, { "id": 34, "seek": 16660, "start": 172.6, "end": 176.6, "text": "落到你身上就是找得到,問得動,不用你動手整理。", "tokens": [ 50665, 30848, 4511, 2166, 19847, 5708, 5620, 25085, 5916, 4511, 11, 11361, 5916, 12572, 11, 24384, 2166, 12572, 11389, 27662, 13876, 1543, 50865 ], "temperature": 0, "avg_logprob": -0.06698141911233119, "compression_ratio": 1.3008130081300813, "no_speech_prob": 2.0075172427191923e-11, "source": "whisper" }, { "id": 35, "seek": 16660, "start": 176.6, "end": 184.6, "text": "知識分三類收,常見問題,文檔,知識頁,收進來自動打標籤,找的時候按標籤篩。", "tokens": [ 50865, 6498, 43143, 6627, 10960, 40092, 18681, 11, 11279, 10470, 17197, 11, 17174, 40741, 242, 11, 6498, 43143, 8313, 223, 11, 18681, 18214, 3763, 9722, 12572, 12467, 46299, 22113, 97, 11, 25085, 20643, 26613, 46299, 22113, 97, 20878, 102, 1543, 51265 ], "temperature": 0, "avg_logprob": -0.06698141911233119, "compression_ratio": 1.3008130081300813, "no_speech_prob": 2.0075172427191923e-11, "source": "whisper" }, { "id": 36, "seek": 16660, "start": 184.6, "end": 189.6, "text": "公司那幾百份沒人看得完的文檔,扔進去就是個隨時能問的庫。", "tokens": [ 51265, 13545, 32981, 4184, 23575, 31906, 36266, 4913, 4035, 4200, 5916, 14128, 1546, 17174, 40741, 242, 11, 3416, 242, 18214, 6734, 5620, 3338, 48890, 6611, 8225, 11361, 1546, 6346, 104, 1543, 51515 ], "temperature": 0, "avg_logprob": -0.06698141911233119, "compression_ratio": 1.3008130081300813, "no_speech_prob": 2.0075172427191923e-11, "source": "whisper" }, { "id": 37, "seek": 18960, "start": 189.6, "end": 198.6, "text": "考場把一整套資料丟進去,考前只管問,三千多次被人搬回去自己改,不是只有點心為觀的。", "tokens": [ 50365, 26504, 18523, 33175, 16075, 2257, 27662, 1881, 245, 35380, 33404, 940, 253, 18214, 6734, 11, 26504, 8945, 14003, 23131, 11361, 11, 10960, 20787, 6392, 9487, 23238, 4035, 22220, 105, 49925, 17645, 34490, 11, 7296, 35244, 8216, 7945, 6344, 25179, 1546, 1543, 50815 ], "temperature": 0, "avg_logprob": -0.09764463837082321, "compression_ratio": 1.1085714285714285, "no_speech_prob": 2.050483220716881e-11, "source": "whisper", "text_original": "考政黨把一整套資料丟進去,考前只管問,三千多次被人搬回去自己改,不是只有點心為觀的。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 38, "seek": 18960, "start": 198.6, "end": 203.6, "text": "聽著,像直接把文檔丟給聊天框,可兩邊根本不是一回事。", "tokens": [ 50815, 21524, 19382, 11, 12760, 43297, 16075, 17174, 40741, 242, 940, 253, 17798, 40096, 6135, 21416, 228, 11, 4429, 16313, 14901, 31337, 8802, 7296, 2257, 8350, 6973, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.09764463837082321, "compression_ratio": 1.1085714285714285, "no_speech_prob": 2.050483220716881e-11, "source": "whisper" }, { "id": 39, "seek": 20360, "start": 203.6, "end": 206.6, "text": "差就差在誰替你管這堆東西。", "tokens": [ 50365, 21679, 3111, 21679, 3581, 20757, 9531, 123, 2166, 23131, 2664, 10726, 228, 26978, 1543, 50515 ], "temperature": 0, "avg_logprob": -0.12637673786708287, "compression_ratio": 1.1993006993006994, "no_speech_prob": 1.981577575638216e-11, "source": "whisper" }, { "id": 40, "seek": 20360, "start": 206.6, "end": 211.6, "text": "你平時把文檔丟給聊天框?餵一次答一次,那次答完他就忘了。", "tokens": [ 50515, 2166, 16716, 6611, 16075, 17174, 40741, 242, 940, 253, 17798, 40096, 6135, 21416, 228, 30, 22217, 113, 27505, 28223, 27505, 11, 4184, 9487, 28223, 14128, 5000, 3111, 26677, 2289, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.12637673786708287, "compression_ratio": 1.1993006993006994, "no_speech_prob": 1.981577575638216e-11, "source": "whisper" }, { "id": 41, "seek": 20360, "start": 211.6, "end": 214.6, "text": "下次新開一局,之前餵的全白費。", "tokens": [ 50765, 35829, 12560, 8949, 2257, 34703, 11, 32442, 22217, 113, 1546, 11319, 13558, 50071, 1543, 50915 ], "temperature": 0, "avg_logprob": -0.12637673786708287, "compression_ratio": 1.1993006993006994, "no_speech_prob": 1.981577575638216e-11, "source": "whisper" }, { "id": 42, "seek": 20360, "start": 214.6, "end": 217.6, "text": "Week Nora這邊十幾種格式都能吃。", "tokens": [ 50915, 4360, 916, 45741, 22821, 20145, 23575, 16516, 30921, 27584, 7182, 8225, 10123, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.12637673786708287, "compression_ratio": 1.1993006993006994, "no_speech_prob": 1.981577575638216e-11, "source": "whisper" }, { "id": 43, "seek": 20360, "start": 217.6, "end": 222.6, "text": "PDF,文檔,表格,幻燈片,網頁存檔,甚至圖片和思維導圖。", "tokens": [ 51065, 17349, 37, 11, 17174, 40741, 242, 11, 17571, 30921, 11, 3509, 119, 25874, 16668, 11, 26352, 8313, 223, 39781, 40741, 242, 11, 17556, 20844, 2523, 244, 16668, 12565, 8870, 9261, 255, 25647, 2523, 244, 1543, 51315 ], "temperature": 0, "avg_logprob": -0.12637673786708287, "compression_ratio": 1.1993006993006994, "no_speech_prob": 1.981577575638216e-11, "source": "whisper", "text_original": "PDF,文檔,表格,幻登片,網頁存檔,甚至圖片和思維導圖。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 44, "seek": 20360, "start": 222.6, "end": 227.6, "text": "你手裡是什麼樣,扔進去就是什麼樣,不用先轉格式。", "tokens": [ 51315, 2166, 11389, 11066, 35851, 6180, 11, 3416, 242, 18214, 6734, 5620, 7598, 6180, 11, 24384, 10108, 32739, 30921, 27584, 1543, 51565 ], "temperature": 0, "avg_logprob": -0.12637673786708287, "compression_ratio": 1.1993006993006994, "no_speech_prob": 1.981577575638216e-11, "source": "whisper" }, { "id": 45, "seek": 22760, "start": 227.6, "end": 235.6, "text": "整個文件夾拖進去,原來什麼目錄結構進來還是什麼樣,攢了多年的資料堆,原樣搬進來就行。", "tokens": [ 50365, 27662, 3338, 17174, 20485, 1787, 122, 6852, 244, 18214, 6734, 11, 19683, 3763, 7598, 11386, 9567, 226, 17144, 43362, 18214, 3763, 25583, 7598, 6180, 11, 5917, 95, 2289, 6392, 34128, 35380, 33404, 10726, 228, 11, 19683, 6180, 22220, 105, 18214, 3763, 3111, 8082, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.10832816362380981, "compression_ratio": 1.2857142857142858, "no_speech_prob": 2.0204306977467112e-11, "source": "whisper" }, { "id": 46, "seek": 22760, "start": 235.6, "end": 240.6, "text": "找出來的那一段,要是不對,你可以直接改,改完它自己重新建一遍索引。", "tokens": [ 50765, 25085, 29741, 1546, 4184, 2257, 28427, 11, 4275, 1541, 15779, 11, 42766, 43297, 34490, 11, 34490, 14128, 11284, 17645, 12624, 12560, 34157, 2257, 3330, 235, 7239, 1543, 51015 ], "temperature": 0, "avg_logprob": -0.10832816362380981, "compression_ratio": 1.2857142857142858, "no_speech_prob": 2.0204306977467112e-11, "source": "whisper", "text_original": "找出來的那一段,要是不對,你可以直接改,改完它自己重新建一遍所以。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 47, "seek": 22760, "start": 240.6, "end": 242.6, "text": "酷是越用越準的。", "tokens": [ 51015, 16911, 115, 1541, 25761, 9254, 25761, 28220, 1546, 1543, 51115 ], "temperature": 0, "avg_logprob": -0.10832816362380981, "compression_ratio": 1.2857142857142858, "no_speech_prob": 2.0204306977467112e-11, "source": "whisper" }, { "id": 48, "seek": 22760, "start": 242.6, "end": 251.6, "text": "問答能掛到自己網站上,官方管這個叫域名白名單,就是只有你認的網站能接,還能限流。", "tokens": [ 51115, 11361, 28223, 8225, 6900, 249, 4511, 17645, 26352, 34155, 5708, 11, 31929, 9249, 23131, 6287, 19855, 14356, 110, 15940, 13558, 15940, 24227, 11, 5620, 35244, 2166, 22041, 1546, 26352, 34155, 8225, 14468, 11, 7824, 8225, 43446, 27854, 1543, 51565 ], "temperature": 0, "avg_logprob": -0.10832816362380981, "compression_ratio": 1.2857142857142858, "no_speech_prob": 2.0204306977467112e-11, "source": "whisper", "text_original": "問答能掛到自己網站上,官方管這個叫育名白名單,就是只有你認的網站能接,還能限流。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 49, "seek": 25160, "start": 251.6, "end": 257.6, "text": "開店的,把問答掛到店裡,重複的問題不用一遍遍答,也能接進群裡。", "tokens": [ 50365, 8949, 22469, 1546, 11, 16075, 11361, 28223, 6900, 249, 4511, 22469, 11066, 11, 12624, 164, 97, 229, 1546, 17197, 24384, 2257, 3330, 235, 3330, 235, 28223, 11, 6404, 8225, 14468, 18214, 7248, 97, 11066, 1543, 50665 ], "temperature": 0, "avg_logprob": -0.08381174866264389, "compression_ratio": 1.3578947368421053, "no_speech_prob": 1.8525901299137892e-11, "source": "whisper" }, { "id": 50, "seek": 25160, "start": 257.6, "end": 263.6, "text": "企業微信,飛書,釘釘,微信,都行,在群裡問一句,答案直接回在群裡。", "tokens": [ 50665, 7384, 223, 27119, 39152, 17665, 11, 34629, 29801, 11, 940, 223, 940, 223, 11, 39152, 17665, 11, 7182, 8082, 11, 3581, 7248, 97, 11066, 11361, 2257, 34592, 11, 28223, 28899, 43297, 8350, 3581, 7248, 97, 11066, 1543, 50965 ], "temperature": 0, "avg_logprob": -0.08381174866264389, "compression_ratio": 1.3578947368421053, "no_speech_prob": 1.8525901299137892e-11, "source": "whisper", "text_original": "企業微信,飛書,丁丁,微信,都行,在群裡問一句,答案直接回在群裡。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 51, "seek": 25160, "start": 263.6, "end": 265.6, "text": "帶新人的群最合適。", "tokens": [ 50965, 32379, 12560, 4035, 1546, 7248, 97, 8661, 14245, 3330, 102, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.08381174866264389, "compression_ratio": 1.3578947368421053, "no_speech_prob": 1.8525901299137892e-11, "source": "whisper" }, { "id": 52, "seek": 25160, "start": 265.6, "end": 267.6, "text": "同樣的問題不用答三遍。", "tokens": [ 51065, 13089, 6180, 1546, 17197, 24384, 28223, 10960, 3330, 235, 1543, 51165 ], "temperature": 0, "avg_logprob": -0.08381174866264389, "compression_ratio": 1.3578947368421053, "no_speech_prob": 1.8525901299137892e-11, "source": "whisper" }, { "id": 53, "seek": 25160, "start": 267.6, "end": 273.6, "text": "刷到有用的網頁,瀏覽器擴展,選中就存進庫,官方管這個叫第二大腦。", "tokens": [ 51165, 2437, 115, 4511, 2412, 9254, 1546, 26352, 8313, 223, 11, 163, 222, 237, 2888, 121, 34386, 13167, 112, 43491, 11, 21625, 5975, 3111, 39781, 18214, 6346, 104, 11, 31929, 9249, 23131, 6287, 19855, 19693, 3582, 21184, 99, 1543, 51465 ], "temperature": 0, "avg_logprob": -0.08381174866264389, "compression_ratio": 1.3578947368421053, "no_speech_prob": 1.8525901299137892e-11, "source": "whisper" }, { "id": 54, "seek": 25160, "start": 273.6, "end": 277.6, "text": "手機上還有小程序,接著問,晚上躺床上,接著查。", "tokens": [ 51465, 11389, 17543, 5708, 15569, 7322, 29649, 6346, 237, 11, 14468, 19382, 11361, 11, 50157, 17514, 118, 6346, 232, 5708, 11, 14468, 19382, 42623, 1543, 51665 ], "temperature": 0, "avg_logprob": -0.08381174866264389, "compression_ratio": 1.3578947368421053, "no_speech_prob": 1.8525901299137892e-11, "source": "whisper" }, { "id": 55, "seek": 27760, "start": 277.6, "end": 285.6, "text": "多提一句向量這個詞,你可以理解成給每段文字編的號,機器靠這個號按意思找東西。", "tokens": [ 50365, 6392, 20949, 2257, 34592, 24282, 26748, 6287, 6716, 252, 11, 42766, 13876, 17278, 11336, 17798, 23664, 28427, 17174, 22381, 12608, 101, 1546, 18616, 11, 17543, 34386, 5363, 254, 6287, 18616, 26613, 16697, 25085, 26978, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.10198521258226081, "compression_ratio": 1.2640845070422535, "no_speech_prob": 2.1947950401557392e-11, "source": "whisper" }, { "id": 56, "seek": 27760, "start": 285.6, "end": 291.6, "text": "號存在專門的數據庫裡,就是它放資料的那個倉庫,七八家可選。", "tokens": [ 50765, 18616, 39781, 3581, 46072, 37232, 1546, 30622, 36841, 6346, 104, 11066, 11, 5620, 11284, 12744, 35380, 33404, 1546, 20754, 2122, 231, 6346, 104, 11, 29920, 33453, 5155, 4429, 21625, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.10198521258226081, "compression_ratio": 1.2640845070422535, "no_speech_prob": 2.1947950401557392e-11, "source": "whisper" }, { "id": 57, "seek": 27760, "start": 291.6, "end": 297.6, "text": "文檔和模型都能留在自己機器上,這是你自己的庫,不用每次重新喂一遍。", "tokens": [ 51065, 17174, 40741, 242, 12565, 41908, 39823, 7182, 8225, 24456, 3581, 17645, 17543, 34386, 5708, 11, 16600, 2166, 17645, 1546, 6346, 104, 11, 24384, 23664, 9487, 12624, 12560, 35142, 2257, 3330, 235, 1543, 51365 ], "temperature": 0, "avg_logprob": -0.10198521258226081, "compression_ratio": 1.2640845070422535, "no_speech_prob": 2.1947950401557392e-11, "source": "whisper" }, { "id": 58, "seek": 27760, "start": 297.6, "end": 302.6, "text": "官網演示裡,拿競品分析報告當例子,一句話讓它整理。", "tokens": [ 51365, 31929, 26352, 31382, 25696, 11066, 11, 24351, 11957, 114, 30246, 6627, 7360, 238, 17803, 16846, 13118, 17797, 7626, 11, 2257, 34592, 11103, 21195, 11284, 27662, 13876, 1543, 51615 ], "temperature": 0, "avg_logprob": -0.10198521258226081, "compression_ratio": 1.2640845070422535, "no_speech_prob": 2.1947950401557392e-11, "source": "whisper" }, { "id": 59, "seek": 30260, "start": 302.6, "end": 310.6, "text": "還展示從三個知識庫,照回十幾條相關片段,再調數據看板,讀本季度銷售指標。", "tokens": [ 50365, 7824, 43491, 25696, 21068, 10960, 3338, 6498, 43143, 6346, 104, 11, 32150, 8350, 20145, 23575, 31106, 15106, 14899, 16668, 28427, 11, 8623, 32943, 30622, 36841, 4200, 43664, 11, 7422, 222, 8802, 8052, 96, 13127, 42041, 115, 17198, 106, 25922, 46299, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.09751918023092705, "compression_ratio": 1.2469135802469136, "no_speech_prob": 2.2136863522814743e-11, "source": "whisper" }, { "id": 60, "seek": 30260, "start": 310.6, "end": 316.6, "text": "這段是官網上的演示,不是我實測出來的。擺出來是給你看它想幹什麼活。", "tokens": [ 50765, 2664, 28427, 1541, 31929, 26352, 5708, 1546, 31382, 25696, 11, 7296, 1654, 10376, 9592, 105, 29741, 1546, 1543, 13167, 118, 29741, 1541, 17798, 16529, 11284, 7093, 29932, 7598, 25956, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.09751918023092705, "compression_ratio": 1.2469135802469136, "no_speech_prob": 2.2136863522814743e-11, "source": "whisper" }, { "id": 61, "seek": 30260, "start": 316.6, "end": 320.6, "text": "東西是好東西,上手到底要折騰到什麼程度。", "tokens": [ 51065, 26978, 1541, 2131, 26978, 11, 5708, 11389, 33883, 4275, 7235, 246, 38168, 108, 4511, 7598, 29649, 13127, 1543, 51265 ], "temperature": 0, "avg_logprob": -0.09751918023092705, "compression_ratio": 1.2469135802469136, "no_speech_prob": 2.2136863522814743e-11, "source": "whisper" }, { "id": 62, "seek": 30260, "start": 320.6, "end": 323.6, "text": "先說最實在的上手路,把它下下來。", "tokens": [ 51265, 10108, 4622, 8661, 10376, 3581, 1546, 5708, 11389, 9567, 226, 11, 42061, 4438, 27769, 1543, 51415 ], "temperature": 0, "avg_logprob": -0.09751918023092705, "compression_ratio": 1.2469135802469136, "no_speech_prob": 2.2136863522814743e-11, "source": "whisper", "text_original": "先說最實在的上手錄,把它下下來。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 63, "seek": 32360, "start": 323.6, "end": 330.6, "text": "改一行配置,敲兩句,啟動命令,瀏覽器打開,就是個能用的網站。", "tokens": [ 50365, 34490, 2257, 8082, 38846, 34719, 11, 7017, 110, 16313, 34592, 11, 3284, 253, 12572, 25236, 49061, 11, 163, 222, 237, 2888, 121, 34386, 12467, 8949, 11, 5620, 3338, 8225, 9254, 1546, 26352, 34155, 1543, 50715 ], "temperature": 0, "avg_logprob": -0.07924580723984437, "compression_ratio": 1.3514492753623188, "no_speech_prob": 1.8717357525011025e-11, "source": "whisper" }, { "id": 64, "seek": 32360, "start": 330.6, "end": 332.6, "text": "前後四行,多一行都沒有。", "tokens": [ 50715, 8945, 5661, 19425, 8082, 11, 6392, 2257, 8082, 7182, 6963, 1543, 50815 ], "temperature": 0, "avg_logprob": -0.07924580723984437, "compression_ratio": 1.3514492753623188, "no_speech_prob": 1.8717357525011025e-11, "source": "whisper" }, { "id": 65, "seek": 32360, "start": 332.6, "end": 334.6, "text": "前置條件只有兩樣。", "tokens": [ 50815, 8945, 34719, 31106, 20485, 35244, 16313, 6180, 1543, 50915 ], "temperature": 0, "avg_logprob": -0.07924580723984437, "compression_ratio": 1.3514492753623188, "no_speech_prob": 1.8717357525011025e-11, "source": "whisper" }, { "id": 66, "seek": 32360, "start": 334.6, "end": 339.6, "text": "一種把軟件連環境一起打包好,敲兩句就能啟動的工具。", "tokens": [ 50915, 2257, 16516, 16075, 14454, 253, 20485, 36252, 40063, 108, 47885, 29567, 12467, 23305, 2131, 11, 7017, 110, 16313, 34592, 3111, 8225, 3284, 253, 12572, 1546, 23323, 39806, 1543, 51165 ], "temperature": 0, "avg_logprob": -0.07924580723984437, "compression_ratio": 1.3514492753623188, "no_speech_prob": 1.8717357525011025e-11, "source": "whisper" }, { "id": 67, "seek": 32360, "start": 339.6, "end": 342.6, "text": "再加一個把代碼拉下來的小工具。", "tokens": [ 51165, 8623, 9990, 8990, 16075, 19105, 16337, 120, 29717, 27769, 1546, 7322, 23323, 39806, 1543, 51315 ], "temperature": 0, "avg_logprob": -0.07924580723984437, "compression_ratio": 1.3514492753623188, "no_speech_prob": 1.8717357525011025e-11, "source": "whisper" }, { "id": 68, "seek": 32360, "start": 342.6, "end": 344.6, "text": "嫌敲命令麻煩。", "tokens": [ 51315, 5196, 240, 7017, 110, 25236, 49061, 38999, 17867, 102, 1543, 51415 ], "temperature": 0, "avg_logprob": -0.07924580723984437, "compression_ratio": 1.3514492753623188, "no_speech_prob": 1.8717357525011025e-11, "source": "whisper", "text_original": "閒敲命令麻煩。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 69, "seek": 32360, "start": 344.6, "end": 347.6, "text": "官網還有個騰訊雲一鍵部署的按鈕。", "tokens": [ 51415, 31929, 26352, 15569, 3338, 38168, 108, 5396, 232, 6306, 110, 2257, 41668, 113, 13470, 16469, 110, 1546, 26613, 41597, 243, 1543, 51565 ], "temperature": 0, "avg_logprob": -0.07924580723984437, "compression_ratio": 1.3514492753623188, "no_speech_prob": 1.8717357525011025e-11, "source": "whisper" }, { "id": 70, "seek": 32360, "start": 347.6, "end": 349.6, "text": "一鍵這兩個字,是人家自己寫的。", "tokens": [ 51565, 2257, 41668, 113, 2664, 34623, 22381, 11, 1541, 4035, 5155, 17645, 4510, 104, 1546, 1543, 51665 ], "temperature": 0, "avg_logprob": -0.07924580723984437, "compression_ratio": 1.3514492753623188, "no_speech_prob": 1.8717357525011025e-11, "source": "whisper" }, { "id": 71, "seek": 34960, "start": 349.6, "end": 354.6, "text": "但這條走的是騰訊雲上的服務,不是跑在你自己電腦上。", "tokens": [ 50365, 8395, 2664, 31106, 9575, 24620, 38168, 108, 5396, 232, 6306, 110, 5708, 1546, 27408, 39517, 11, 7296, 32585, 3581, 2166, 17645, 20545, 21184, 99, 5708, 1543, 50615 ], "temperature": 0, "avg_logprob": -0.05680797411047894, "compression_ratio": 1.5376344086021505, "no_speech_prob": 1.7039127053752878e-11, "source": "whisper" }, { "id": 72, "seek": 34960, "start": 354.6, "end": 356.6, "text": "要不要錢,怎麼算。", "tokens": [ 50615, 4275, 11962, 29483, 11, 11614, 19497, 1543, 50715 ], "temperature": 0, "avg_logprob": -0.05680797411047894, "compression_ratio": 1.5376344086021505, "no_speech_prob": 1.7039127053752878e-11, "source": "whisper" }, { "id": 73, "seek": 34960, "start": 356.6, "end": 358.6, "text": "你得上那頁自己看。", "tokens": [ 50715, 2166, 5916, 5708, 4184, 8313, 223, 17645, 4200, 1543, 50815 ], "temperature": 0, "avg_logprob": -0.05680797411047894, "compression_ratio": 1.5376344086021505, "no_speech_prob": 1.7039127053752878e-11, "source": "whisper" }, { "id": 74, "seek": 34960, "start": 358.6, "end": 362.6, "text": "前置條件那一欄裡,一個顯存數字都沒提。", "tokens": [ 50815, 8945, 34719, 31106, 20485, 4184, 2257, 5988, 226, 11066, 11, 8990, 46605, 39781, 30622, 22381, 7182, 4913, 20949, 1543, 51015 ], "temperature": 0, "avg_logprob": -0.05680797411047894, "compression_ratio": 1.5376344086021505, "no_speech_prob": 1.7039127053752878e-11, "source": "whisper" }, { "id": 75, "seek": 34960, "start": 362.6, "end": 364.6, "text": "因為模型不在它身上跑。", "tokens": [ 51015, 11471, 41908, 39823, 1960, 3581, 11284, 19847, 5708, 32585, 1543, 51115 ], "temperature": 0, "avg_logprob": -0.05680797411047894, "compression_ratio": 1.5376344086021505, "no_speech_prob": 1.7039127053752878e-11, "source": "whisper" }, { "id": 76, "seek": 34960, "start": 364.6, "end": 369.6, "text": "你想全離線,顯卡的賬得算在你選的那個本地模型頭上。", "tokens": [ 51115, 38386, 11319, 32662, 31699, 11, 46605, 32681, 1546, 13352, 105, 5916, 19497, 3581, 2166, 21625, 1546, 20754, 8802, 10928, 41908, 39823, 19921, 5708, 1543, 51365 ], "temperature": 0, "avg_logprob": -0.05680797411047894, "compression_ratio": 1.5376344086021505, "no_speech_prob": 1.7039127053752878e-11, "source": "whisper" }, { "id": 77, "seek": 34960, "start": 369.6, "end": 374.6, "text": "走本地模型工具,也就是一個能在自己電腦上跑開源模型的工具。", "tokens": [ 51365, 9575, 8802, 10928, 41908, 39823, 23323, 39806, 11, 6404, 5620, 8990, 8225, 3581, 17645, 20545, 21184, 99, 5708, 32585, 8949, 47402, 41908, 39823, 1546, 23323, 39806, 1543, 51615 ], "temperature": 0, "avg_logprob": -0.05680797411047894, "compression_ratio": 1.5376344086021505, "no_speech_prob": 1.7039127053752878e-11, "source": "whisper" }, { "id": 78, "seek": 34960, "start": 374.6, "end": 378.6, "text": "文檔不出自己電腦,模型也在自己電腦上跑。", "tokens": [ 51615, 17174, 40741, 242, 1960, 7781, 17645, 20545, 21184, 99, 11, 41908, 39823, 6404, 3581, 17645, 20545, 21184, 99, 5708, 32585, 1543, 51815 ], "temperature": 0, "avg_logprob": -0.05680797411047894, "compression_ratio": 1.5376344086021505, "no_speech_prob": 1.7039127053752878e-11, "source": "whisper" }, { "id": 79, "seek": 37860, "start": 378.6, "end": 382.6, "text": "這條路官方顯明了支持,線上模型支持20來家。", "tokens": [ 50365, 2664, 31106, 24658, 31929, 9249, 46605, 11100, 2289, 35488, 11, 31699, 5708, 41908, 39823, 35488, 2009, 3763, 5155, 1543, 50565 ], "temperature": 0, "avg_logprob": -0.08321176256452288, "compression_ratio": 1.2925170068027212, "no_speech_prob": 1.84826685206696e-11, "source": "whisper" }, { "id": 80, "seek": 37860, "start": 382.6, "end": 384.6, "text": "國產幾家也在裡面。", "tokens": [ 50565, 8053, 33299, 23575, 5155, 6404, 3581, 32399, 1543, 50665 ], "temperature": 0, "avg_logprob": -0.08321176256452288, "compression_ratio": 1.2925170068027212, "no_speech_prob": 1.84826685206696e-11, "source": "whisper" }, { "id": 81, "seek": 37860, "start": 384.6, "end": 389.6, "text": "不過默認接的是線上模型,全離線,要自己換成本地那套。", "tokens": [ 50665, 28193, 6173, 246, 22041, 14468, 24620, 31699, 5708, 41908, 39823, 11, 11319, 32662, 31699, 11, 4275, 17645, 36338, 11336, 8802, 10928, 4184, 1881, 245, 1543, 50915 ], "temperature": 0, "avg_logprob": -0.08321176256452288, "compression_ratio": 1.2925170068027212, "no_speech_prob": 1.84826685206696e-11, "source": "whisper" }, { "id": 82, "seek": 37860, "start": 389.6, "end": 391.6, "text": "別裝完就以為已經斷網了。", "tokens": [ 50915, 16158, 44453, 14128, 3111, 3588, 6344, 25338, 4307, 115, 26352, 2289, 1543, 51015 ], "temperature": 0, "avg_logprob": -0.08321176256452288, "compression_ratio": 1.2925170068027212, "no_speech_prob": 1.84826685206696e-11, "source": "whisper" }, { "id": 83, "seek": 37860, "start": 391.6, "end": 393.6, "text": "還有一條壓根不用自己裝的路,", "tokens": [ 51015, 15569, 2257, 31106, 45964, 31337, 24384, 17645, 44453, 1546, 24658, 11, 51115 ], "temperature": 0, "avg_logprob": -0.08321176256452288, "compression_ratio": 1.2925170068027212, "no_speech_prob": 1.84826685206696e-11, "source": "whisper" }, { "id": 84, "seek": 37860, "start": 393.6, "end": 398.6, "text": "騰訊把這套東西做成了微信的一個對話開放平台。", "tokens": [ 51115, 38168, 108, 5396, 232, 16075, 2664, 1881, 245, 26978, 10907, 11336, 2289, 39152, 17665, 4184, 14901, 1546, 8990, 2855, 11103, 8949, 12744, 16716, 15433, 1543, 51365 ], "temperature": 0, "avg_logprob": -0.08321176256452288, "compression_ratio": 1.2925170068027212, "no_speech_prob": 1.84826685206696e-11, "source": "whisper", "text_original": "騰訊把這套東西做成了微信那邊的一個對話開放平台。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 85, "seek": 37860, "start": 398.6, "end": 403.6, "text": "傳知識就能用,不用碰命令行,也就是不用敲那幾行字。", "tokens": [ 51365, 40852, 6498, 43143, 3111, 8225, 9254, 11, 24384, 16337, 108, 25236, 49061, 8082, 11, 6404, 5620, 24384, 7017, 110, 4184, 23575, 8082, 22381, 1543, 51615 ], "temperature": 0, "avg_logprob": -0.08321176256452288, "compression_ratio": 1.2925170068027212, "no_speech_prob": 1.84826685206696e-11, "source": "whisper" }, { "id": 86, "seek": 40360, "start": 403.6, "end": 407.6, "text": "連裝都不想裝的人走這條就行。", "tokens": [ 50365, 36252, 44453, 7182, 1960, 7093, 44453, 29979, 9575, 2664, 31106, 3111, 8082, 1543, 50565 ], "temperature": 0, "avg_logprob": -0.12257520207818949, "compression_ratio": 1.182608695652174, "no_speech_prob": 2.1073540482086095e-11, "source": "whisper" }, { "id": 87, "seek": 40360, "start": 407.6, "end": 411.6, "text": "官方另外準備了一個輕量版,有單獨的安裝配方。", "tokens": [ 50565, 31929, 9249, 26202, 36241, 2289, 8990, 34801, 26748, 42096, 11, 2412, 24227, 163, 18304, 1546, 16206, 44453, 38846, 9249, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.12257520207818949, "compression_ratio": 1.182608695652174, "no_speech_prob": 2.1073540482086095e-11, "source": "whisper" }, { "id": 88, "seek": 40360, "start": 411.6, "end": 416.6, "text": "能省多少?跟完整版差在哪?說明文檔裡沒寫,我也不替它吹。", "tokens": [ 50765, 8225, 2862, 223, 41492, 30, 9678, 14128, 27662, 42096, 21679, 3581, 17028, 30, 4622, 11100, 17174, 40741, 242, 11066, 4913, 8082, 11, 27761, 1960, 9531, 123, 11284, 2392, 117, 1543, 51015 ], "temperature": 0, "avg_logprob": -0.12257520207818949, "compression_ratio": 1.182608695652174, "no_speech_prob": 2.1073540482086095e-11, "source": "whisper", "text_original": "能省多少?跟完整版差在哪?說明文檔裡沒行,我也不替它吹。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 89, "seek": 40360, "start": 416.6, "end": 418.6, "text": "等官方把話說全再講。", "tokens": [ 51015, 10187, 31929, 9249, 16075, 11103, 4622, 11319, 8623, 11932, 1543, 51115 ], "temperature": 0, "avg_logprob": -0.12257520207818949, "compression_ratio": 1.182608695652174, "no_speech_prob": 2.1073540482086095e-11, "source": "whisper" }, { "id": 90, "seek": 40360, "start": 418.6, "end": 422.6, "text": "自己攢了多年的電子書堆,整個目錄拖進去。", "tokens": [ 51115, 17645, 5917, 240, 2289, 6392, 34128, 20545, 7626, 29801, 10726, 228, 11, 27662, 3338, 11386, 24658, 6852, 244, 18214, 6734, 1543, 51315 ], "temperature": 0, "avg_logprob": -0.12257520207818949, "compression_ratio": 1.182608695652174, "no_speech_prob": 2.1073540482086095e-11, "source": "whisper", "text_original": "自己攒了多年的電子書堆,整個目路拖進去。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 91, "seek": 42260, "start": 422.6, "end": 425.6, "text": "出門路上想查就查。", "tokens": [ 50365, 7781, 37232, 24658, 5708, 7093, 42623, 3111, 42623, 1543, 50515 ], "temperature": 0, "avg_logprob": -0.0884734755712198, "compression_ratio": 1.2841726618705036, "no_speech_prob": 1.973652491438216e-11, "source": "whisper" }, { "id": 92, "seek": 42260, "start": 425.6, "end": 427.6, "text": "好處說完了,那代價呢?", "tokens": [ 50515, 2131, 39289, 4622, 41665, 11, 4184, 19105, 45441, 6240, 30, 50615 ], "temperature": 0, "avg_logprob": -0.0884734755712198, "compression_ratio": 1.2841726618705036, "no_speech_prob": 1.973652491438216e-11, "source": "whisper" }, { "id": 93, "seek": 42260, "start": 427.6, "end": 430.6, "text": "白送的是軟件,運維是你自己的。", "tokens": [ 50615, 13558, 29309, 24620, 14454, 253, 20485, 11, 32772, 9261, 255, 1541, 2166, 17645, 1546, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.0884734755712198, "compression_ratio": 1.2841726618705036, "no_speech_prob": 1.973652491438216e-11, "source": "whisper" }, { "id": 94, "seek": 42260, "start": 430.6, "end": 438.6, "text": "官方原話寫在安全提示裡,裝在內部網自己用,別直接掛到公網,掛出去會漏東西。", "tokens": [ 50765, 31929, 9249, 19683, 11103, 4510, 104, 3581, 45466, 20949, 25696, 11066, 11, 44453, 3581, 28472, 13470, 26352, 17645, 9254, 11, 16158, 43297, 6900, 249, 4511, 13545, 26352, 11, 6900, 249, 48973, 6236, 14065, 237, 26978, 1543, 51165 ], "temperature": 0, "avg_logprob": -0.0884734755712198, "compression_ratio": 1.2841726618705036, "no_speech_prob": 1.973652491438216e-11, "source": "whisper" }, { "id": 95, "seek": 42260, "start": 438.6, "end": 440.6, "text": "這是他自己寫的,不是我嚇唬你。", "tokens": [ 51165, 16600, 5000, 17645, 4510, 104, 1546, 11, 7296, 1654, 29593, 229, 17198, 105, 2166, 1543, 51265 ], "temperature": 0, "avg_logprob": -0.0884734755712198, "compression_ratio": 1.2841726618705036, "no_speech_prob": 1.973652491438216e-11, "source": "whisper" }, { "id": 96, "seek": 42260, "start": 440.6, "end": 448.6, "text": "登錄驗證是後來才補上的,最早的版本連登錄都沒有,這條也寫在同一段安全提示裡。", "tokens": [ 51265, 25874, 9567, 226, 24023, 245, 11823, 231, 1541, 5661, 3763, 18888, 8083, 250, 5708, 1546, 11, 8661, 21176, 1546, 42096, 8802, 36252, 25874, 9567, 226, 7182, 6963, 11, 2664, 31106, 6404, 4510, 104, 3581, 13089, 2257, 28427, 45466, 20949, 25696, 11066, 1543, 51665 ], "temperature": 0, "avg_logprob": -0.0884734755712198, "compression_ratio": 1.2841726618705036, "no_speech_prob": 1.973652491438216e-11, "source": "whisper" }, { "id": 97, "seek": 44860, "start": 448.6, "end": 452.6, "text": "門鎖是後裝的,早下手的人裸奔過一陣。", "tokens": [ 50365, 37232, 165, 236, 244, 1541, 5661, 44453, 1546, 11, 21176, 4438, 11389, 29979, 8083, 116, 1881, 242, 8816, 2257, 8842, 96, 1543, 50565 ], "temperature": 0, "avg_logprob": -0.058375719736313676, "compression_ratio": 1.267741935483871, "no_speech_prob": 1.7036764360378598e-11, "source": "whisper" }, { "id": 98, "seek": 44860, "start": 452.6, "end": 458.6, "text": "讓AI在你機器上動手腳的地方,官方一直在收緊,默認關著,手動才能開。", "tokens": [ 50565, 21195, 48698, 3581, 2166, 17543, 34386, 5708, 12572, 11389, 49595, 1546, 30146, 11, 31929, 9249, 34448, 3581, 18681, 30172, 11, 6173, 246, 22041, 14899, 19382, 11, 11389, 12572, 18888, 8225, 8949, 1543, 50865 ], "temperature": 0, "avg_logprob": -0.058375719736313676, "compression_ratio": 1.267741935483871, "no_speech_prob": 1.7036764360378598e-11, "source": "whisper" }, { "id": 99, "seek": 44860, "start": 458.6, "end": 463.6, "text": "剛到0.8版,能動手腳的路又砍掉一條,這點我支持官方。", "tokens": [ 50865, 16940, 4511, 15, 13, 23, 42096, 11, 8225, 12572, 11389, 49595, 1546, 24658, 17047, 23230, 235, 29327, 2257, 31106, 11, 2664, 8216, 1654, 1593, 242, 31929, 9249, 1543, 51115 ], "temperature": 0, "avg_logprob": -0.058375719736313676, "compression_ratio": 1.267741935483871, "no_speech_prob": 1.7036764360378598e-11, "source": "whisper", "text_original": "剛到0.8版,能動手腳的路又砍掉一條,這點我佔官方。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 100, "seek": 44860, "start": 463.6, "end": 472.6, "text": "角色權限分四檔,還有審計日誌,多工作區,邀請鏈,動了哪份文檔,誰動的,日誌裡都記著。", "tokens": [ 51115, 29389, 17673, 40212, 43446, 6627, 19425, 40741, 242, 11, 15569, 4510, 102, 30114, 6890, 3549, 234, 11, 6392, 41315, 44526, 11, 3023, 222, 16302, 3549, 234, 11, 12572, 2289, 17028, 36266, 17174, 40741, 242, 11, 20757, 12572, 1546, 11, 6890, 3549, 234, 11066, 7182, 16958, 19382, 1543, 51565 ], "temperature": 0, "avg_logprob": -0.058375719736313676, "compression_ratio": 1.267741935483871, "no_speech_prob": 1.7036764360378598e-11, "source": "whisper", "text_original": "角色權限分四檔,還有審計日誌,多工作區,邀請誌,動了哪份文檔,誰動的,日誌裡都記著。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 101, "seek": 44860, "start": 472.6, "end": 476.6, "text": "一個人的資料庫,被迫先讀懂一套管幾百人的規矩。", "tokens": [ 51565, 8990, 4035, 1546, 35380, 33404, 6346, 104, 11, 23238, 3316, 104, 10108, 7422, 222, 29624, 2257, 1881, 245, 23131, 23575, 31906, 4035, 1546, 2888, 237, 5881, 102, 1543, 51765 ], "temperature": 0, "avg_logprob": -0.058375719736313676, "compression_ratio": 1.267741935483871, "no_speech_prob": 1.7036764360378598e-11, "source": "whisper" }, { "id": 102, "seek": 47660, "start": 476.6, "end": 486.6, "text": "裝起來不是一個小軟件,是一套服務。存編號的數據庫要配,放文件的,排任務的也得配上,拼起來才能轉。", "tokens": [ 50365, 44453, 21670, 7296, 8990, 7322, 14454, 253, 20485, 11, 1541, 2257, 1881, 245, 27408, 39517, 1543, 39781, 12608, 101, 18616, 1546, 30622, 36841, 6346, 104, 4275, 38846, 11, 12744, 17174, 20485, 1546, 11, 44647, 26443, 39517, 1546, 6404, 5916, 38846, 5708, 11, 6852, 120, 21670, 18888, 8225, 32739, 1543, 50865 ], "temperature": 0, "avg_logprob": -0.060890127111364295, "compression_ratio": 1.25, "no_speech_prob": 1.9906272810676917e-11, "source": "whisper" }, { "id": 103, "seek": 47660, "start": 486.6, "end": 490.6, "text": "愛折騰的人覺得是自由,怕麻煩的人看到就頭大。", "tokens": [ 50865, 15157, 7235, 246, 38168, 108, 29979, 11274, 1541, 45233, 11, 21164, 38999, 17867, 102, 29979, 18032, 3111, 19921, 3582, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.060890127111364295, "compression_ratio": 1.25, "no_speech_prob": 1.9906272810676917e-11, "source": "whisper" }, { "id": 104, "seek": 47660, "start": 490.6, "end": 496.6, "text": "防火牆和訪問控制自己配,補丁自己打,管這套東西的人就是你自己。", "tokens": [ 51065, 35863, 26171, 6935, 228, 12565, 5396, 103, 11361, 11, 48707, 25491, 17645, 38846, 11, 8083, 250, 940, 223, 17645, 12467, 11, 23131, 2664, 1881, 245, 26978, 29979, 5620, 2166, 17645, 1543, 51365 ], "temperature": 0, "avg_logprob": -0.060890127111364295, "compression_ratio": 1.25, "no_speech_prob": 1.9906272810676917e-11, "source": "whisper", "text_original": "防火牆和訪問,控制自己配,補丁自己打,管這套東西的人就是你自己。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 105, "seek": 49660, "start": 496.6, "end": 502.6, "text": "半夜出問題,搖不到騰訊的人,只能自己爬起來看。", "tokens": [ 50365, 30018, 30124, 7781, 17197, 11, 22220, 244, 35728, 38168, 108, 5396, 232, 29979, 11, 14003, 8225, 17645, 8164, 105, 21670, 4200, 1543, 50665 ], "temperature": 0, "avg_logprob": -0.10115188540834369, "compression_ratio": 1.0794701986754967, "no_speech_prob": 1.9076079257884793e-11, "source": "whisper" }, { "id": 106, "seek": 49660, "start": 502.6, "end": 506.6, "text": "三百多條問題還開著,將近三百個沒合的改動請求。", "tokens": [ 50665, 10960, 31906, 6392, 31106, 17197, 7824, 8949, 19382, 11, 34168, 17463, 10960, 31906, 3338, 4913, 14245, 1546, 34490, 12572, 16302, 32718, 1543, 50865 ], "temperature": 0, "avg_logprob": -0.10115188540834369, "compression_ratio": 1.0794701986754967, "no_speech_prob": 1.9076079257884793e-11, "source": "whisper" }, { "id": 107, "seek": 49660, "start": 506.6, "end": 508.6, "text": "活躍是活躍,坑也不少。", "tokens": [ 50865, 25956, 17514, 235, 1541, 25956, 17514, 235, 11, 14872, 239, 6404, 1960, 15686, 1543, 50965 ], "temperature": 0, "avg_logprob": -0.10115188540834369, "compression_ratio": 1.0794701986754967, "no_speech_prob": 1.9076079257884793e-11, "source": "whisper" }, { "id": 108, "seek": 50860, "start": 508.6, "end": 512.6, "text": "所以我的判斷是,文檔多到,翻不動的人值得折騰。", "tokens": [ 50365, 7239, 14200, 2437, 97, 4307, 115, 1541, 11, 17174, 40741, 242, 6392, 4511, 11, 42716, 1960, 12572, 29979, 40242, 5916, 7235, 246, 38168, 108, 1543, 50565 ], "temperature": 0, "avg_logprob": -0.05837276134085148, "compression_ratio": 1.1818181818181819, "no_speech_prob": 1.9002518308885996e-11, "source": "whisper" }, { "id": 109, "seek": 50860, "start": 512.6, "end": 515.6, "text": "手裡只有幾十份文檔,真沒必要折騰。", "tokens": [ 50565, 11389, 11066, 35244, 23575, 20145, 36266, 17174, 40741, 242, 11, 6303, 4913, 28531, 4275, 7235, 246, 38168, 108, 1543, 50715 ], "temperature": 0, "avg_logprob": -0.05837276134085148, "compression_ratio": 1.1818181818181819, "no_speech_prob": 1.9002518308885996e-11, "source": "whisper" }, { "id": 110, "seek": 50860, "start": 515.6, "end": 519.6, "text": "先從微信那邊免部署那條路,順手再自己裝。", "tokens": [ 50715, 10108, 21068, 39152, 17665, 4184, 14901, 2347, 235, 13470, 16469, 110, 4184, 31106, 6973, 11, 8313, 228, 11389, 8623, 17645, 44453, 1543, 50915 ], "temperature": 0, "avg_logprob": -0.05837276134085148, "compression_ratio": 1.1818181818181819, "no_speech_prob": 1.9002518308885996e-11, "source": "whisper", "text_original": "先從微信那邊免部署那條事,順手再自己裝。", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" }, { "id": 111, "seek": 50860, "start": 519.6, "end": 522.6, "text": "買不買賬,取決於你手裡文檔有多少。", "tokens": [ 50915, 31001, 1960, 31001, 13352, 105, 11, 29436, 33540, 19488, 2166, 11389, 11066, 17174, 40741, 242, 2412, 41492, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.05837276134085148, "compression_ratio": 1.1818181818181819, "no_speech_prob": 1.9002518308885996e-11, "source": "whisper" }, { "id": 112, "seek": 52260, "start": 522.6, "end": 525.6, "text": "你們那幾百份文檔不用再一份份翻了。", "tokens": [ 50365, 19891, 4184, 23575, 31906, 36266, 17174, 40741, 242, 24384, 8623, 2257, 36266, 36266, 42716, 2289, 1543, 50515 ], "temperature": 0, "avg_logprob": -0.094345703125, "compression_ratio": 1.3758620689655172, "no_speech_prob": 1.88606751744258e-11, "source": "whisper" }, { "id": 113, "seek": 52260, "start": 525.6, "end": 530.6, "text": "騰訊開源了,一套能把它們變成隨時能問的知識庫。", "tokens": [ 50515, 38168, 108, 5396, 232, 8949, 47402, 2289, 11, 2257, 1881, 245, 8225, 42061, 4623, 26327, 11336, 48890, 6611, 8225, 11361, 1546, 6498, 43143, 6346, 104, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.094345703125, "compression_ratio": 1.3758620689655172, "no_speech_prob": 1.88606751744258e-11, "source": "whisper" }, { "id": 114, "seek": 52260, "start": 530.6, "end": 535.6, "text": "文檔還能不出公司,這類能自己裝起來的國產開源工具。", "tokens": [ 50765, 17174, 40741, 242, 7824, 8225, 1960, 7781, 13545, 32981, 11, 2664, 40092, 8225, 17645, 44453, 21670, 1546, 8053, 33299, 8949, 47402, 23323, 39806, 1543, 51015 ], "temperature": 0, "avg_logprob": -0.094345703125, "compression_ratio": 1.3758620689655172, "no_speech_prob": 1.88606751744258e-11, "source": "whisper" }, { "id": 115, "seek": 52260, "start": 535.6, "end": 538.6, "text": "我會一個個替你試過再講。", "tokens": [ 51015, 1654, 6236, 8990, 3338, 9531, 123, 2166, 22099, 8816, 8623, 11932, 1543, 51165 ], "temperature": 0, "avg_logprob": -0.094345703125, "compression_ratio": 1.3758620689655172, "no_speech_prob": 1.88606751744258e-11, "source": "whisper" }, { "id": 116, "seek": 52260, "start": 538.6, "end": 540.6, "text": "包括那些其實跑不起來的。", "tokens": [ 51165, 41828, 4184, 13824, 14139, 32585, 1960, 21670, 1546, 1543, 51265 ], "temperature": 0, "avg_logprob": -0.094345703125, "compression_ratio": 1.3758620689655172, "no_speech_prob": 1.88606751744258e-11, "source": "whisper" }, { "id": 117, "seek": 52260, "start": 540.6, "end": 545.6, "text": "最後問你一句,公司那堆文檔你是想自己裝一套慢慢問,", "tokens": [ 51265, 20578, 11361, 2166, 2257, 34592, 11, 13545, 32981, 4184, 10726, 228, 17174, 40741, 242, 32526, 7093, 17645, 44453, 2257, 1881, 245, 41439, 11361, 11, 51515 ], "temperature": 0, "avg_logprob": -0.094345703125, "compression_ratio": 1.3758620689655172, "no_speech_prob": 1.88606751744258e-11, "source": "whisper" }, { "id": 118, "seek": 52260, "start": 545.6, "end": 547.6, "text": "還是直接傳知識用現成的?", "tokens": [ 51515, 25583, 43297, 40852, 6498, 43143, 9254, 9581, 11336, 1546, 30, 51615 ], "temperature": 0, "avg_logprob": -0.094345703125, "compression_ratio": 1.3758620689655172, "no_speech_prob": 1.88606751744258e-11, "source": "whisper" }, { "id": 119, "seek": 52260, "start": 547.6, "end": 549.6, "text": "裝過的講講卡在哪一步?", "tokens": [ 51615, 44453, 8816, 1546, 11932, 11932, 32681, 3581, 17028, 2257, 31429, 30, 51715 ], "temperature": 0, "avg_logprob": -0.094345703125, "compression_ratio": 1.3758620689655172, "no_speech_prob": 1.88606751744258e-11, "source": "whisper" }, { "id": 120, "seek": 54960, "start": 549.6, "end": 554.581375, "text": "留言區的評論區報個一或者二,你選哪個?", "tokens": [ 50365, 2862, 231, 31127, 232, 1546, 39115, 29340, 44526, 17803, 3338, 2257, 31148, 11217, 11, 2166, 21625, 17028, 3338, 30, 50615 ], "temperature": 0, "avg_logprob": -0.2902561101046475, "compression_ratio": 0.819672131147541, "no_speech_prob": 1.6695636192998187e-11, "source": "whisper", "text_original": "眉莊的評論區報個一或者二,你選哪個?", "correction_applied": true, "correction_source": "llm_correct_a_presegmentation" } ], "transcript_selection": { "schema": "videonote.transcript_selection.v1", "status": "SELECTED", "requested_language": "zh", "duration": 554.581375, "reference_path": null, "reference_kind": null, "reference_language": null, "selected_source": "whisper_segmented", "decision_reason": "source_unavailable_or_untrusted_whisper_passed", "whisper_rejected": false, "source_profile": { "segment_count": 0, "duration_coverage": 0.0, "usable_full": false, "usable_partial": false, "hallucination_region_count": 0, "hallucination_regions": [], "language": { "requested": "zh", "matched": false } }, "whisper_profile": { "segment_count": 121, "multilingual": false, "first_start": 0.0, "final_end": 554.581, "duration_coverage": 1.0, "maximum_gap_seconds": 0.0, "inverted_count": 0, "language": { "requested": "zh", "matched": true, "latin": 46, "han": 2460, "kana": 0, "hangul": 0 }, "hallucination_region_count": 0, "hallucination_regions": [], "usable_full": true, "usable_partial": true }, "comparison": { "window_seconds": 30.0, "compared_window_count": 0, "median_score": 0.0, "divergent_window_count": 0, "divergent_ratio": 0.0, "divergent_windows": [] }, "timestamp_alignment": { "applied": false, "estimated_offset_seconds": 0.0, "matched_segment_count": 0, "match_scores": [] }, "caption_timing_repair": null, "filled_source_profile": null, "whisper_duration_guard": { "media_duration": 554.581, "rejected_segment_count": 0, "rejected_segments": [] } }, "presegmentation_correction": { "applied": true, "correction_count": 18, "deterministic_fix_count": 0 } }