{ "text": "大家好,我是小莫頭。OpenAI近期又發布了一系列的新模型。其中有一款是專門做實時的同傳或實時翻譯。似乎在AI時代要做同聲傳譯簡單了許多。本期視頻呢,我們就來實際的體驗一下這款實時翻譯的模型。看看在同聲傳譯上它的質量究竟如何。\n我們首先來看個演示吧。我將模型封裝了一個小小的Chrome插件。AIN把它命名為了Bubblefish。插件的作用是真聽當前所瀏覽的頁面中的音頻。將音頻轉換成期望的目標語言。這個轉換當然是語音層面的轉換。我就以過去我所發布的一期視頻為例。這是一個中文視頻。我希望將目標語言翻譯成英文。點擊Start。\n開始堅定當前瀏覽頁面中的音頻數據。我們就來播放一段語音。看看它的轉換效果。\n我介紹一下它的開放的開放的設計。\n在一段影片中,我介紹了閉設設計。如果您興趣,您可以再看這影片。我會介紹一下它的開放。\n在CloDesign,我們可以建立的開放。\n我們可以建立的開放。\n建立開放的開放。\n我們可以建立開放的開放。\n在這影片中,我們會介紹一下開放的開放。\n開放的開放。\n在這影片中,我們會介紹一下開放的開放。\n開放開放的開放。\n在這影片中,我們會介紹開放開放開放的開放。\nOPDesign是今天的星星星。\n這個開放的開放。\n這個開放的開放是現在非常有名的。\n在短時間,它們會集中了約20,000星星星。\n我會放開開放的開放。\n影片的開放在影片的描述。\n興趣的朋友,請給大家看看。\n影片中的影片。\n這個開放是一個第一次開放開放的開放。\n我們會看看它的主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主\n接下来我们来做两件事情首先呢是把这个模型做一番清晰的介绍\n接下来再来看看我们用它来构建的这款Chrome插件究竟是怎么回事\nOpenAI实际上是在5月7日在实施API下放了三个新模型\nGPT Real-Time 2\nGPT Real-Time Translate\nGPT Real-Time Whisper\n在官方的公告中列出了这三款模型\nGPT Real-Time Translate是我们今天分享的主角\n它做一件事就是把语音同步的翻译\n70多种输入语言自动检测\n13种输出语言任选\n中文英文西班牙语葡萄牙语法语德语等等\n机费层面来讲0.034美元每分钟\n看一小时英文播客大约2美元\n比人类同传便宜两个数量级\n但它跟之前的通用语音模型加Prompt让它翻译并不是一回事\n根据OpenAI的Cookbook\n他们说得很清晰\n这个模型是用同传演员的语料训练的\n具备两个特征\n第一\n它会等到上下文够了才开口\n不会人家一开口\n它就抢着翻\n这对语序差异大的语言特别重要\n英语和日语\n英语和中文\n位置动词位置不一样\n等不够上下文\n就乱翻\n这是不行的\n如果大家实际的测试一下应该能感觉到相较于原始的YouTube的音频\n实时翻译的音频在段句子之间可能会有一定的延迟\n这也是在上下文等待中产生的效果\n第二个特征是它能边收音边图异音\n它并不是speech to text转文字\n再翻译\n再通过text speech这种三段的拼接\n而是端到端的speech to speech\n做到真正的低延迟连续语音\nAPI形态则有real-time endpoint\n浏览器用webrtc\n服务端用websocket\nOpenAI Cookbook有一个demo\n就是浏览器tab翻译\n目前我所分享的这个目录呢\n就是刚才咱们介绍的这个demo\n我会将链接也放在视频描述中\n那么在刚才我们的演示中\n所用到的chrome插件呢\n就是现在我所分享的这个代码仓库\n这是一个chrome插件\n它包含了三个组件\n一个是service worker\n一个是offscreen document\n还包含了一部分content script\n总共是一个不到600行的type script代码\n非常简单\n感兴趣的朋友可以客套下来自己尝试一下\n大家要做的就是几步\n首先呢\n客套项目\n我们通过npm install\nnpm run build\n完成安装与打包\n接下来来到chrome浏览器\n管理插件\n在插件这里\n首先要在右上角\n启用开发者模式\n然后点击load unpacked\n在open real-time translate目录当中\n选择dist\n这个目录呢\n就是打包后生成的目录\n要使用这个插件也很简单\n在扩展这里点击它\n第一次使用的时候呢\n需要配个api key\n大家去寻一个自己的open-air的api key\n配置进来保存\n然后任意打开一个标签页\n比如播放一段YouTube视频\n那我们现在呢\n再来做一番演示\n比如我想将我这个语音呢\n实时的转换成日语\n来听听效果吧\n音質的には\n丹丸コーポレートアイデンティティは持たず\n既存のコーポレートアイデンティティ機能や能力を活用して\nopen source のエコシステムにある\n非常に豊かなスキルマッケージを統合し\nそれによって\nクラウドデザインと似た\nあるいはそれ以上に\n機能統合の豊かなopen source 精神を構築しました\nopen design は\n現時点のopen source エコシステムで\n非常に人気のある\nいくつかのプロジェクトを統合しています\n例えば\n花玉デザインや\n北角師匠のPTTスキルなど\nopen course design なども含まれます\n細心的朋友可能注意到了\n剛才我將視頻暂停了\n但是語音輸出还在持續了一小段時間\n剛才我們在介绍模型的时候也聊到这一塊\n模型会收集一段必要的上下文再做翻譯\n因此在这个翻譯和語音輸出上\n就会有一点小小的延遲\n这也是它的特点吧\n来看看開銷情况吧\n我大概前前後後在視頻录制前的準備到\n视频中的实际分享大概\n在語音的輸出上花了可能有半小时\n那么目前的开销是0.9亿美元\n那现在就对咱们今天分享做一番简单的总结\n首先呢\n这个chrome插件\n目前只是抓取的浏览器tab中的音频\n像zoom桌面客户端\nspotify桌面\n本地视频播放器等等呢\n是抓不到的\n不过同样的逻辑或者是代码呢\n可以应用到不同的应用场景中\n比如可以覆盖到我们刚才介绍的这些桌面应用\n至于具体的技术瓶颈在哪里\n是否操作系统支持音频的捕获呢\n这个就是我们在技术或工程层面需要去考虑的了\n第二点是GPT Real-Time Translate\n这个实时翻译模型\n它不支持自定义的提示词\n数语表\n配音选择\n如果你的场景对专有名词敏感\n它会按一般的翻译习惯来处理\n另外一点很重要的是\nSession没有关掉\n它会一直跑\n一直击费\n一小时大概2美元\n四小时大概就是8美元\n因此呢\n大家用完记得要Stop\n在插件层面\n大家就记得点击Stop Translating\nRestop\n如果大家不再使用了呢\n点击Stop\n停掉它\n否则呢\n这个Session一直会开着\n整个扩展从想法到能用\n目前在AEA时代编写起来非常的快\n给到它足够的参考文档\nAPI手册\n很快就搞定了\n你可以把它分叉下来\n加自己想要的功能\n比如把字幕导出到SRT\n加Hockey切换语言\n把意音导进OBS做直播同传等等\n这都是直接的扩展点\nOpenAI把同传从需要专业人员的服务\n变成了一个API调用\n这会不会是同传这个职业\n也是跨语言内容消费方式的一个明显的转折点呢\n大家怎么看呢\n那感兴趣的朋友\n或者在这方面已经有应用经验的朋友\n欢迎在评论区留言\n那今天的分享就到这里\n感谢大家收看\n我们下期再见", "language": "zh", "duration": 565.963188, "segments": [ { "id": 0, "seek": 0, "start": 0.0, "end": 23.0, "text": "大家好,我是小莫頭。OpenAI近期又發布了一系列的新模型。其中有一款是專門做實時的同傳或實時翻譯。似乎在AI時代要做同聲傳譯簡單了許多。本期視頻呢,我們就來實際的體驗一下這款實時翻譯的模型。看看在同聲傳譯上它的質量究竟如何。", "tokens": [ 50365, 15248, 11, 15914, 7322, 31127, 104, 19921, 1543, 45569, 48698, 17463, 16786, 17047, 14637, 34688, 2289, 2257, 25368, 43338, 1546, 12560, 41908, 39823, 1543, 9572, 5975, 32241, 48798, 1541, 46072, 37232, 10907, 10376, 6611, 1546, 13089, 40852, 19780, 10376, 6611, 42716, 11823, 107, 1543, 7384, 120, 2930, 236, 3581, 48698, 6611, 19105, 4275, 10907, 13089, 29454, 40852, 11823, 107, 31995, 2289, 40914, 6392, 1543, 8802, 16786, 27333, 29940, 6240, 11, 5884, 3111, 3763, 10376, 34837, 1546, 23987, 24023, 245, 8861, 2664, 48798, 10376, 6611, 42716, 11823, 107, 1546, 41908, 39823, 1543, 22884, 3581, 13089, 29454, 40852, 11823, 107, 5708, 45224, 43450, 26748, 44704, 11957, 253, 43526, 1543, 51515 ], "temperature": 0, "avg_logprob": -0.14519108952702703, "compression_ratio": 1.24609375, "no_speech_prob": 9.316946519843938e-12 }, { "id": 1, "seek": 2300, "start": 23.0, "end": 52.0, "text": "我們首先來看個演示吧。我將模型封裝了一個小小的Chrome插件。AIN把它命名為了Bubblefish。插件的作用是真聽當前所瀏覽的頁面中的音頻。將音頻轉換成期望的目標語言。這個轉換當然是語音層面的轉換。我就以過去我所發布的一期視頻為例。這是一個中文視頻。我希望將目標語言翻譯成英文。點擊Start。", "tokens": [ 50365, 5884, 36490, 3763, 4200, 3338, 31382, 25696, 6062, 1543, 1654, 34168, 41908, 39823, 1530, 223, 44453, 2289, 8990, 7322, 7322, 1546, 11644, 423, 11673, 240, 20485, 1543, 32, 1464, 42061, 25236, 15940, 6344, 2289, 33, 836, 638, 11608, 1543, 11673, 240, 20485, 1546, 11914, 9254, 1541, 6303, 21524, 13118, 8945, 5966, 163, 222, 237, 2888, 121, 1546, 8313, 223, 8833, 5975, 1546, 18034, 29940, 1543, 34168, 18034, 29940, 32739, 36338, 11336, 16786, 22694, 1546, 11386, 46299, 31348, 12009, 1543, 6287, 32739, 36338, 21707, 1541, 31348, 18034, 9636, 97, 8833, 1546, 32739, 36338, 1543, 22020, 3588, 36910, 1654, 5966, 14637, 34688, 1546, 2257, 16786, 27333, 29940, 6344, 17797, 1543, 16600, 8990, 5975, 17174, 27333, 29940, 1543, 1654, 29955, 34168, 11386, 46299, 31348, 12009, 42716, 11823, 107, 11336, 27869, 17174, 1543, 8216, 37813, 33996, 1543, 51815 ], "temperature": 0, "avg_logprob": -0.14572673685410442, "compression_ratio": 1.3094462540716612, "no_speech_prob": 1.1805723065805296e-11 }, { "id": 2, "seek": 5300, "start": 53.0, "end": 61.0, "text": "開始堅定當前瀏覽頁面中的音頻數據。我們就來播放一段語音。看看它的轉換效果。", "tokens": [ 50365, 21017, 10726, 227, 12088, 13118, 8945, 163, 222, 237, 2888, 121, 8313, 223, 8833, 5975, 1546, 18034, 29940, 30622, 36841, 1543, 5884, 3111, 3763, 49993, 12744, 2257, 28427, 31348, 18034, 1543, 22884, 45224, 32739, 36338, 43076, 9319, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.5989006278441125, "compression_ratio": 1.433673469387755, "no_speech_prob": 2.5439138642835424e-11 }, { "id": 3, "seek": 5300, "start": 61.0, "end": 66.0, "text": "我介紹一下它的開放的開放的設計。", "tokens": [ 50765, 1654, 30312, 38193, 8861, 45224, 8949, 12744, 1546, 8949, 12744, 1546, 39035, 30114, 1543, 51015 ], "temperature": 0, "avg_logprob": -0.5989006278441125, "compression_ratio": 1.433673469387755, "no_speech_prob": 2.5439138642835424e-11 }, { "id": 4, "seek": 5300, "start": 66.0, "end": 79.0, "text": "在一段影片中,我介紹了閉設設計。如果您興趣,您可以再看這影片。我會介紹一下它的開放。", "tokens": [ 51015, 3581, 2257, 28427, 25373, 5975, 11, 1654, 30312, 38193, 2289, 5196, 231, 39035, 39035, 30114, 1543, 13119, 23414, 44089, 39835, 11, 23414, 6723, 8623, 4200, 2664, 25373, 1543, 1654, 6236, 30312, 38193, 8861, 45224, 8949, 12744, 1543, 51665 ], "temperature": 0, "avg_logprob": -0.5989006278441125, "compression_ratio": 1.433673469387755, "no_speech_prob": 2.5439138642835424e-11 }, { "id": 5, "seek": 7900, "start": 79.0, "end": 83.0, "text": "在CloDesign,我們可以建立的開放。", "tokens": [ 50365, 3581, 34, 752, 29543, 788, 11, 5884, 6723, 34157, 24409, 1546, 8949, 12744, 1543, 50565 ], "temperature": 0, "avg_logprob": -0.9972749328613282, "compression_ratio": 2.078125, "no_speech_prob": 1.2335711341293631e-09 }, { "id": 6, "seek": 7900, "start": 83.0, "end": 85.0, "text": "我們可以建立的開放。", "tokens": [ 50565, 5884, 6723, 34157, 24409, 1546, 8949, 12744, 1543, 50665 ], "temperature": 0, "avg_logprob": -0.9972749328613282, "compression_ratio": 2.078125, "no_speech_prob": 1.2335711341293631e-09 }, { "id": 7, "seek": 7900, "start": 85.0, "end": 87.0, "text": "建立開放的開放。", "tokens": [ 50665, 34157, 24409, 8949, 12744, 1546, 8949, 12744, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.9972749328613282, "compression_ratio": 2.078125, "no_speech_prob": 1.2335711341293631e-09 }, { "id": 8, "seek": 7900, "start": 87.0, "end": 89.0, "text": "我們可以建立開放的開放。", "tokens": [ 50765, 5884, 6723, 34157, 24409, 8949, 12744, 1546, 8949, 12744, 1543, 50865 ], "temperature": 0, "avg_logprob": -0.9972749328613282, "compression_ratio": 2.078125, "no_speech_prob": 1.2335711341293631e-09 }, { "id": 9, "seek": 8900, "start": 89.0, "end": 91.0, "text": "在這影片中,我們會介紹一下開放的開放。", "tokens": [ 50365, 3581, 2664, 25373, 5975, 11, 5884, 6236, 30312, 38193, 8861, 8949, 12744, 1546, 8949, 12744, 1543, 50465 ], "temperature": 0, "avg_logprob": -0.6671465465000698, "compression_ratio": 2.1714285714285713, "no_speech_prob": 1.2767810142477742e-09 }, { "id": 10, "seek": 8900, "start": 91.0, "end": 93.0, "text": "開放的開放。", "tokens": [ 50465, 8949, 12744, 1546, 8949, 12744, 1543, 50565 ], "temperature": 0, "avg_logprob": -0.6671465465000698, "compression_ratio": 2.1714285714285713, "no_speech_prob": 1.2767810142477742e-09 }, { "id": 11, "seek": 8900, "start": 93.0, "end": 95.0, "text": "在這影片中,我們會介紹一下開放的開放。", "tokens": [ 50565, 3581, 2664, 25373, 5975, 11, 5884, 6236, 30312, 38193, 8861, 8949, 12744, 1546, 8949, 12744, 1543, 50665 ], "temperature": 0, "avg_logprob": -0.6671465465000698, "compression_ratio": 2.1714285714285713, "no_speech_prob": 1.2767810142477742e-09 }, { "id": 12, "seek": 8900, "start": 95.0, "end": 97.0, "text": "開放開放的開放。", "tokens": [ 50665, 8949, 12744, 8949, 12744, 1546, 8949, 12744, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.6671465465000698, "compression_ratio": 2.1714285714285713, "no_speech_prob": 1.2767810142477742e-09 }, { "id": 13, "seek": 9700, "start": 97.0, "end": 99.0, "text": "在這影片中,我們會介紹開放開放開放的開放。", "tokens": [ 50365, 3581, 2664, 25373, 5975, 11, 5884, 6236, 30312, 38193, 8949, 12744, 8949, 12744, 8949, 12744, 1546, 8949, 12744, 1543, 50465 ], "temperature": 0, "avg_logprob": -0.44516630985277117, "compression_ratio": 2.8873873873873874, "no_speech_prob": 2.086617857666173e-11 }, { "id": 14, "seek": 9700, "start": 99.0, "end": 101.0, "text": "OPDesign是今天的星星星。", "tokens": [ 50465, 12059, 29543, 788, 1541, 12074, 1546, 20682, 20682, 20682, 1543, 50565 ], "temperature": 0, "avg_logprob": -0.44516630985277117, "compression_ratio": 2.8873873873873874, "no_speech_prob": 2.086617857666173e-11 }, { "id": 15, "seek": 9700, "start": 101.0, "end": 103.0, "text": "這個開放的開放。", "tokens": [ 50565, 6287, 8949, 12744, 1546, 8949, 12744, 1543, 50665 ], "temperature": 0, "avg_logprob": -0.44516630985277117, "compression_ratio": 2.8873873873873874, "no_speech_prob": 2.086617857666173e-11 }, { "id": 16, "seek": 9700, "start": 103.0, "end": 105.0, "text": "這個開放的開放是現在非常有名的。", "tokens": [ 50665, 6287, 8949, 12744, 1546, 8949, 12744, 1541, 12648, 14392, 2412, 15940, 1546, 1543, 50765 ], "temperature": 0, "avg_logprob": -0.44516630985277117, "compression_ratio": 2.8873873873873874, "no_speech_prob": 2.086617857666173e-11 }, { "id": 17, "seek": 9700, "start": 105.0, "end": 108.0, "text": "在短時間,它們會集中了約20,000星星星。", "tokens": [ 50765, 3581, 5881, 255, 20788, 11, 11284, 4623, 6236, 26020, 5975, 2289, 30889, 2009, 11, 1360, 20682, 20682, 20682, 1543, 50915 ], "temperature": 0, "avg_logprob": -0.44516630985277117, "compression_ratio": 2.8873873873873874, "no_speech_prob": 2.086617857666173e-11 }, { "id": 18, "seek": 9700, "start": 108.0, "end": 109.0, "text": "我會放開開放的開放。", "tokens": [ 50915, 1654, 6236, 12744, 8949, 8949, 12744, 1546, 8949, 12744, 1543, 50965 ], "temperature": 0, "avg_logprob": -0.44516630985277117, "compression_ratio": 2.8873873873873874, "no_speech_prob": 2.086617857666173e-11 }, { "id": 19, "seek": 9700, "start": 109.0, "end": 111.0, "text": "影片的開放在影片的描述。", "tokens": [ 50965, 25373, 1546, 8949, 12744, 3581, 25373, 1546, 11673, 237, 3316, 108, 1543, 51065 ], "temperature": 0, "avg_logprob": -0.44516630985277117, "compression_ratio": 2.8873873873873874, "no_speech_prob": 2.086617857666173e-11 }, { "id": 20, "seek": 9700, "start": 111.0, "end": 113.0, "text": "興趣的朋友,請給大家看看。", "tokens": [ 51065, 44089, 39835, 1546, 19828, 11, 16302, 17798, 6868, 22884, 1543, 51165 ], "temperature": 0, "avg_logprob": -0.44516630985277117, "compression_ratio": 2.8873873873873874, "no_speech_prob": 2.086617857666173e-11 }, { "id": 21, "seek": 11300, "start": 113.0, "end": 115.0, "text": "影片中的影片。", "tokens": [ 50365, 25373, 5975, 1546, 25373, 1543, 50465 ], "temperature": 0, "avg_logprob": -0.15918098757619814, "compression_ratio": 7.180851063829787, "no_speech_prob": 1.5259368421616237e-11 }, { "id": 22, "seek": 11300, "start": 115.0, "end": 118.0, "text": "這個開放是一個第一次開放開放的開放。", "tokens": [ 50465, 6287, 8949, 12744, 1541, 8990, 18049, 9487, 8949, 12744, 8949, 12744, 1546, 8949, 12744, 1543, 50615 ], "temperature": 0, "avg_logprob": -0.15918098757619814, "compression_ratio": 7.180851063829787, "no_speech_prob": 1.5259368421616237e-11 }, { "id": 23, "seek": 11800, "start": 118.0, "end": 148.0, "text": "我們會看看它的主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主要主", "tokens": [ 50365, 5884, 6236, 22884, 45224, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557, 4275, 13557 ], "temperature": 0, "avg_logprob": -0.05992790187955437, "compression_ratio": 16.390243902439025, "no_speech_prob": 1.612140108908644e-11 }, { "id": 24, "seek": 14800, "start": 148.0, "end": 152.98, "text": "接下来我们来做两件事情首先呢是把这个模型做一番清晰的介绍", "tokens": [ 50365, 14468, 4438, 6912, 15003, 6912, 10907, 36257, 49459, 36490, 6240, 1541, 16075, 15368, 41908, 39823, 10907, 2257, 33118, 21784, 5094, 108, 1546, 30312, 10115, 235, 50614 ], "temperature": 0, "avg_logprob": -0.09840050379435221, "compression_ratio": 1.2956810631229236, "no_speech_prob": 6.259126550389738e-11 }, { "id": 25, "seek": 14800, "start": 152.98, "end": 158.44, "text": "接下来再来看看我们用它来构建的这款Chrome插件究竟是怎么回事", "tokens": [ 50614, 14468, 4438, 6912, 8623, 6912, 22884, 15003, 9254, 11284, 6912, 7360, 226, 34157, 1546, 5562, 48798, 11644, 423, 11673, 240, 20485, 44704, 11957, 253, 1541, 15282, 8350, 6973, 50887 ], "temperature": 0, "avg_logprob": -0.09840050379435221, "compression_ratio": 1.2956810631229236, "no_speech_prob": 6.259126550389738e-11 }, { "id": 26, "seek": 14800, "start": 158.44, "end": 164.62, "text": "OpenAI实际上是在5月7日在实施API下放了三个新模型", "tokens": [ 50887, 45569, 48698, 24726, 8842, 227, 5708, 1541, 3581, 20, 6939, 22, 6890, 3581, 24726, 4307, 121, 4715, 40, 4438, 12744, 2289, 10960, 7549, 12560, 41908, 39823, 51196 ], "temperature": 0, "avg_logprob": -0.09840050379435221, "compression_ratio": 1.2956810631229236, "no_speech_prob": 6.259126550389738e-11 }, { "id": 27, "seek": 14800, "start": 164.62, "end": 166.42000000000002, "text": "GPT Real-Time 2", "tokens": [ 51196, 38, 47, 51, 8467, 12, 22233, 568, 51286 ], "temperature": 0, "avg_logprob": -0.09840050379435221, "compression_ratio": 1.2956810631229236, "no_speech_prob": 6.259126550389738e-11 }, { "id": 28, "seek": 14800, "start": 166.42000000000002, "end": 168.14, "text": "GPT Real-Time Translate", "tokens": [ 51286, 38, 47, 51, 8467, 12, 22233, 6531, 17593, 51372 ], "temperature": 0, "avg_logprob": -0.09840050379435221, "compression_ratio": 1.2956810631229236, "no_speech_prob": 6.259126550389738e-11 }, { "id": 29, "seek": 14800, "start": 168.14, "end": 169.6, "text": "GPT Real-Time Whisper", "tokens": [ 51372, 38, 47, 51, 8467, 12, 22233, 41132, 610, 51445 ], "temperature": 0, "avg_logprob": -0.09840050379435221, "compression_ratio": 1.2956810631229236, "no_speech_prob": 6.259126550389738e-11 }, { "id": 30, "seek": 14800, "start": 169.6, "end": 173.9, "text": "在官方的公告中列出了这三款模型", "tokens": [ 51445, 3581, 31929, 9249, 1546, 13545, 16846, 5975, 43338, 7781, 2289, 5562, 10960, 48798, 41908, 39823, 51660 ], "temperature": 0, "avg_logprob": -0.09840050379435221, "compression_ratio": 1.2956810631229236, "no_speech_prob": 6.259126550389738e-11 }, { "id": 31, "seek": 14800, "start": 173.9, "end": 177.52, "text": "GPT Real-Time Translate是我们今天分享的主角", "tokens": [ 51660, 38, 47, 51, 8467, 12, 22233, 6531, 17593, 1541, 15003, 12074, 30855, 1546, 13557, 29389, 51841 ], "temperature": 0, "avg_logprob": -0.09840050379435221, "compression_ratio": 1.2956810631229236, "no_speech_prob": 6.259126550389738e-11 }, { "id": 32, "seek": 17752, "start": 177.52, "end": 181.84, "text": "它做一件事就是把语音同步的翻译", "tokens": [ 50365, 11284, 10907, 2257, 20485, 6973, 5620, 16075, 5233, 255, 18034, 13089, 31429, 1546, 42716, 5233, 239, 50581 ], "temperature": 0, "avg_logprob": -0.1223097394226463, "compression_ratio": 1.2056737588652482, "no_speech_prob": 1.8857854514053862e-11 }, { "id": 33, "seek": 17752, "start": 181.84, "end": 184.38000000000002, "text": "70多种输入语言自动检测", "tokens": [ 50581, 5867, 6392, 39810, 9830, 241, 14028, 5233, 255, 12009, 9722, 34961, 16407, 222, 11038, 233, 50708 ], "temperature": 0, "avg_logprob": -0.1223097394226463, "compression_ratio": 1.2056737588652482, "no_speech_prob": 1.8857854514053862e-11 }, { "id": 34, "seek": 17752, "start": 184.38000000000002, "end": 186.64000000000001, "text": "13种输出语言任选", "tokens": [ 50708, 7668, 39810, 9830, 241, 7781, 5233, 255, 12009, 26443, 2215, 231, 50821 ], "temperature": 0, "avg_logprob": -0.1223097394226463, "compression_ratio": 1.2056737588652482, "no_speech_prob": 1.8857854514053862e-11 }, { "id": 35, "seek": 17752, "start": 186.64000000000001, "end": 190.66000000000003, "text": "中文英文西班牙语葡萄牙语法语德语等等", "tokens": [ 50821, 5975, 17174, 27869, 17174, 16220, 44695, 6935, 247, 5233, 255, 14059, 94, 10674, 226, 6935, 247, 5233, 255, 11148, 5233, 255, 35898, 5233, 255, 36000, 51022 ], "temperature": 0, "avg_logprob": -0.1223097394226463, "compression_ratio": 1.2056737588652482, "no_speech_prob": 1.8857854514053862e-11 }, { "id": 36, "seek": 17752, "start": 190.66000000000003, "end": 194.5, "text": "机费层面来讲0.034美元每分钟", "tokens": [ 51022, 37960, 18464, 117, 9636, 224, 8833, 6912, 39255, 15, 13, 11592, 19, 39433, 23664, 6627, 50064, 51214 ], "temperature": 0, "avg_logprob": -0.1223097394226463, "compression_ratio": 1.2056737588652482, "no_speech_prob": 1.8857854514053862e-11 }, { "id": 37, "seek": 17752, "start": 194.5, "end": 197.22, "text": "看一小时英文播客大约2美元", "tokens": [ 51214, 4200, 2257, 7322, 15729, 27869, 17174, 49993, 32316, 3582, 16853, 99, 17, 39433, 51350 ], "temperature": 0, "avg_logprob": -0.1223097394226463, "compression_ratio": 1.2056737588652482, "no_speech_prob": 1.8857854514053862e-11 }, { "id": 38, "seek": 17752, "start": 197.22, "end": 200.28, "text": "比人类同传便宜两个数量级", "tokens": [ 51350, 11706, 4035, 22113, 119, 13089, 7384, 254, 27364, 2415, 250, 36257, 7549, 33188, 26748, 16853, 100, 51503 ], "temperature": 0, "avg_logprob": -0.1223097394226463, "compression_ratio": 1.2056737588652482, "no_speech_prob": 1.8857854514053862e-11 }, { "id": 39, "seek": 17752, "start": 200.28, "end": 206.66000000000003, "text": "但它跟之前的通用语音模型加Prompt让它翻译并不是一回事", "tokens": [ 51503, 8395, 11284, 9678, 32442, 1546, 19550, 9254, 5233, 255, 18034, 41908, 39823, 9990, 47, 4397, 662, 33650, 11284, 42716, 5233, 239, 3509, 114, 7296, 2257, 8350, 6973, 51822 ], "temperature": 0, "avg_logprob": -0.1223097394226463, "compression_ratio": 1.2056737588652482, "no_speech_prob": 1.8857854514053862e-11 }, { "id": 40, "seek": 20666, "start": 206.66, "end": 209.4, "text": "根据OpenAI的Cookbook", "tokens": [ 50365, 31337, 26075, 106, 45569, 48698, 1546, 34, 1212, 2939, 50502 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 41, "seek": 20666, "start": 209.4, "end": 210.79999999999998, "text": "他们说得很清晰", "tokens": [ 50502, 47911, 8090, 5916, 4563, 21784, 5094, 108, 50572 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 42, "seek": 20666, "start": 210.79999999999998, "end": 214.28, "text": "这个模型是用同传演员的语料训练的", "tokens": [ 50572, 15368, 41908, 39823, 1541, 9254, 13089, 7384, 254, 31382, 3606, 246, 1546, 5233, 255, 33404, 7422, 255, 10115, 225, 1546, 50746 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 43, "seek": 20666, "start": 214.28, "end": 215.62, "text": "具备两个特征", "tokens": [ 50746, 39806, 1787, 229, 36257, 7549, 17682, 2172, 223, 50813 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 44, "seek": 20666, "start": 215.62, "end": 216.72, "text": "第一", "tokens": [ 50813, 18049, 50868 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 45, "seek": 20666, "start": 216.72, "end": 219.04, "text": "它会等到上下文够了才开口", "tokens": [ 50868, 11284, 12949, 10187, 4511, 5708, 4438, 17174, 1787, 253, 2289, 18888, 18937, 18144, 50984 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 46, "seek": 20666, "start": 219.04, "end": 220.54, "text": "不会人家一开口", "tokens": [ 50984, 47928, 4035, 5155, 2257, 18937, 18144, 51059 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 47, "seek": 20666, "start": 220.54, "end": 221.48, "text": "它就抢着翻", "tokens": [ 51059, 11284, 3111, 7235, 95, 20708, 42716, 51106 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 48, "seek": 20666, "start": 221.48, "end": 224.5, "text": "这对语序差异大的语言特别重要", "tokens": [ 51106, 5562, 8713, 5233, 255, 6346, 237, 21679, 5702, 224, 39156, 5233, 255, 12009, 17682, 18453, 24928, 51257 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 49, "seek": 20666, "start": 224.5, "end": 225.94, "text": "英语和日语", "tokens": [ 51257, 27869, 5233, 255, 12565, 6890, 5233, 255, 51329 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 50, "seek": 20666, "start": 225.94, "end": 227.14, "text": "英语和中文", "tokens": [ 51329, 27869, 5233, 255, 12565, 5975, 17174, 51389 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 51, "seek": 20666, "start": 227.14, "end": 229.28, "text": "位置动词位置不一样", "tokens": [ 51389, 11160, 34719, 34961, 5233, 235, 11160, 34719, 1960, 2257, 14496, 51496 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 52, "seek": 20666, "start": 229.28, "end": 230.6, "text": "等不够上下文", "tokens": [ 51496, 10187, 1960, 1787, 253, 5708, 4438, 17174, 51562 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 53, "seek": 20666, "start": 230.6, "end": 231.51999999999998, "text": "就乱翻", "tokens": [ 51562, 3111, 2930, 109, 42716, 51608 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 54, "seek": 20666, "start": 231.51999999999998, "end": 232.44, "text": "这是不行的", "tokens": [ 51608, 27455, 24151, 1546, 51654 ], "temperature": 0, "avg_logprob": -0.11484731858776462, "compression_ratio": 1.260536398467433, "no_speech_prob": 1.7384513967266813e-11 }, { "id": 55, "seek": 23244, "start": 232.44, "end": 239.32, "text": "如果大家实际的测试一下应该能感觉到相较于原始的YouTube的音频", "tokens": [ 50365, 13119, 6868, 24726, 8842, 227, 1546, 11038, 233, 5233, 243, 8861, 44297, 44646, 8225, 9709, 24447, 4511, 15106, 9830, 225, 37732, 19683, 15476, 1546, 3223, 41173, 1546, 18034, 39752, 50709 ], "temperature": 0, "avg_logprob": -0.13761496036610704, "compression_ratio": 1.1476190476190475, "no_speech_prob": 1.394421817596525e-11 }, { "id": 56, "seek": 23244, "start": 239.32, "end": 245.82, "text": "实时翻译的音频在段句子之间可能会有一定的延迟", "tokens": [ 50709, 24726, 15729, 42716, 5233, 239, 1546, 18034, 39752, 3581, 28427, 34592, 7626, 9574, 31685, 16657, 12949, 2412, 24272, 1546, 17527, 114, 3316, 253, 51034 ], "temperature": 0, "avg_logprob": -0.13761496036610704, "compression_ratio": 1.1476190476190475, "no_speech_prob": 1.394421817596525e-11 }, { "id": 57, "seek": 23244, "start": 245.82, "end": 249.34, "text": "这也是在上下文等待中产生的效果", "tokens": [ 51034, 5562, 22021, 3581, 5708, 4438, 17174, 10187, 18390, 5975, 1369, 100, 8244, 1546, 43076, 9319, 51210 ], "temperature": 0, "avg_logprob": -0.13761496036610704, "compression_ratio": 1.1476190476190475, "no_speech_prob": 1.394421817596525e-11 }, { "id": 58, "seek": 23244, "start": 249.34, "end": 254.28, "text": "第二个特征是它能边收音边图异音", "tokens": [ 51210, 19693, 7549, 17682, 2172, 223, 1541, 11284, 8225, 40503, 18681, 18034, 40503, 3919, 122, 5702, 224, 18034, 51457 ], "temperature": 0, "avg_logprob": -0.13761496036610704, "compression_ratio": 1.1476190476190475, "no_speech_prob": 1.394421817596525e-11 }, { "id": 59, "seek": 25428, "start": 254.28, "end": 257.36, "text": "它并不是speech to text转文字", "tokens": [ 50365, 11284, 3509, 114, 7296, 7053, 5023, 281, 2487, 17819, 105, 17174, 22381, 50519 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 60, "seek": 25428, "start": 257.36, "end": 258.16, "text": "再翻译", "tokens": [ 50519, 8623, 42716, 5233, 239, 50559 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 61, "seek": 25428, "start": 258.16, "end": 261.82, "text": "再通过text speech这种三段的拼接", "tokens": [ 50559, 8623, 19550, 16866, 25111, 6218, 5562, 39810, 10960, 28427, 1546, 6852, 120, 14468, 50742 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 62, "seek": 25428, "start": 261.82, "end": 264.8, "text": "而是端到端的speech to speech", "tokens": [ 50742, 11070, 1541, 11957, 107, 4511, 11957, 107, 1546, 7053, 5023, 281, 6218, 50891 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 63, "seek": 25428, "start": 264.8, "end": 267.7, "text": "做到真正的低延迟连续语音", "tokens": [ 50891, 10907, 4511, 6303, 15789, 1546, 41377, 17527, 114, 3316, 253, 3316, 252, 10115, 255, 5233, 255, 18034, 51036 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 64, "seek": 25428, "start": 267.7, "end": 270.54, "text": "API形态则有real-time endpoint", "tokens": [ 51036, 4715, 40, 30900, 3757, 223, 2437, 247, 2412, 9342, 12, 3766, 35795, 51178 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 65, "seek": 25428, "start": 270.54, "end": 272.32, "text": "浏览器用webrtc", "tokens": [ 51178, 11038, 237, 6758, 230, 34386, 9254, 826, 1443, 83, 66, 51267 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 66, "seek": 25428, "start": 272.32, "end": 273.9, "text": "服务端用websocket", "tokens": [ 51267, 27408, 5087, 94, 11957, 107, 9254, 826, 65, 539, 4737, 51346 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 67, "seek": 25428, "start": 273.9, "end": 276.66, "text": "OpenAI Cookbook有一个demo", "tokens": [ 51346, 45569, 48698, 12259, 2939, 2412, 20182, 10730, 78, 51484 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 68, "seek": 25428, "start": 276.66, "end": 278.16, "text": "就是浏览器tab翻译", "tokens": [ 51484, 5620, 11038, 237, 6758, 230, 34386, 83, 455, 42716, 5233, 239, 51559 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 69, "seek": 25428, "start": 278.16, "end": 281.16, "text": "目前我所分享的这个目录呢", "tokens": [ 51559, 39004, 1654, 5966, 30855, 1546, 15368, 11386, 7391, 243, 6240, 51709 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 70, "seek": 25428, "start": 281.16, "end": 283.5, "text": "就是刚才咱们介绍的这个demo", "tokens": [ 51709, 5620, 49160, 18888, 8975, 109, 9497, 30312, 10115, 235, 1546, 15368, 10730, 78, 51826 ], "temperature": 0, "avg_logprob": -0.15524883511700208, "compression_ratio": 1.2096219931271477, "no_speech_prob": 1.5381225806909704e-11 }, { "id": 71, "seek": 28350, "start": 283.5, "end": 285.66, "text": "我会将链接也放在视频描述中", "tokens": [ 50365, 1654, 12949, 45456, 165, 241, 122, 14468, 6404, 12744, 3581, 40656, 39752, 11673, 237, 3316, 108, 5975, 50473 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 72, "seek": 28350, "start": 285.66, "end": 288.34, "text": "那么在刚才我们的演示中", "tokens": [ 50473, 35693, 3581, 49160, 18888, 15003, 1546, 31382, 25696, 5975, 50607 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 73, "seek": 28350, "start": 288.34, "end": 290.14, "text": "所用到的chrome插件呢", "tokens": [ 50607, 5966, 9254, 4511, 1546, 339, 11505, 11673, 240, 20485, 6240, 50697 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 74, "seek": 28350, "start": 290.14, "end": 292.82, "text": "就是现在我所分享的这个代码仓库", "tokens": [ 50697, 5620, 25040, 1654, 5966, 30855, 1546, 15368, 19105, 23230, 223, 1550, 241, 6346, 241, 50831 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 75, "seek": 28350, "start": 292.82, "end": 295.28, "text": "这是一个chrome插件", "tokens": [ 50831, 27455, 20182, 339, 11505, 11673, 240, 20485, 50954 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 76, "seek": 28350, "start": 295.28, "end": 296.86, "text": "它包含了三个组件", "tokens": [ 50954, 11284, 23305, 2392, 104, 2289, 10960, 7549, 10115, 226, 20485, 51033 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 77, "seek": 28350, "start": 296.86, "end": 298.88, "text": "一个是service worker", "tokens": [ 51033, 20182, 1541, 39279, 11346, 51134 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 78, "seek": 28350, "start": 298.88, "end": 301.16, "text": "一个是offscreen document", "tokens": [ 51134, 20182, 1541, 4506, 12439, 4166, 51248 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 79, "seek": 28350, "start": 301.16, "end": 303.7, "text": "还包含了一部分content script", "tokens": [ 51248, 14852, 23305, 2392, 104, 2289, 2257, 32174, 9000, 317, 5755, 51375 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 80, "seek": 28350, "start": 303.7, "end": 307.62, "text": "总共是一个不到600行的type script代码", "tokens": [ 51375, 33440, 15408, 1541, 20182, 35728, 15707, 8082, 1546, 20467, 5755, 19105, 23230, 223, 51571 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 81, "seek": 28350, "start": 307.62, "end": 308.38, "text": "非常简单", "tokens": [ 51571, 14392, 11249, 222, 47446, 51609 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 82, "seek": 28350, "start": 308.38, "end": 311.4, "text": "感兴趣的朋友可以客套下来自己尝试一下", "tokens": [ 51609, 9709, 2347, 112, 39835, 1546, 19828, 6723, 32316, 1881, 245, 4438, 6912, 17645, 1530, 251, 5233, 243, 8861, 51760 ], "temperature": 0, "avg_logprob": -0.11144100889867666, "compression_ratio": 1.2745098039215685, "no_speech_prob": 1.777223680721196e-11 }, { "id": 83, "seek": 31140, "start": 311.4, "end": 313.84, "text": "大家要做的就是几步", "tokens": [ 50365, 6868, 4275, 10907, 1546, 5620, 6336, 254, 31429, 50487 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 84, "seek": 31140, "start": 313.84, "end": 314.44, "text": "首先呢", "tokens": [ 50487, 36490, 6240, 50517 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 85, "seek": 31140, "start": 314.44, "end": 315.14, "text": "客套项目", "tokens": [ 50517, 32316, 1881, 245, 10178, 117, 11386, 50552 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 86, "seek": 31140, "start": 315.14, "end": 317.0, "text": "我们通过npm install", "tokens": [ 50552, 15003, 19550, 16866, 77, 14395, 3625, 50645 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 87, "seek": 31140, "start": 317.0, "end": 318.14, "text": "npm run build", "tokens": [ 50645, 77, 14395, 1190, 1322, 50702 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 88, "seek": 31140, "start": 318.14, "end": 319.44, "text": "完成安装与打包", "tokens": [ 50702, 41509, 16206, 8083, 227, 940, 236, 12467, 23305, 50767 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 89, "seek": 31140, "start": 319.44, "end": 323.02, "text": "接下来来到chrome浏览器", "tokens": [ 50767, 14468, 4438, 6912, 6912, 4511, 339, 11505, 11038, 237, 6758, 230, 34386, 50946 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 90, "seek": 31140, "start": 323.02, "end": 324.35999999999996, "text": "管理插件", "tokens": [ 50946, 23131, 13876, 11673, 240, 20485, 51013 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 91, "seek": 31140, "start": 324.35999999999996, "end": 326.97999999999996, "text": "在插件这里", "tokens": [ 51013, 3581, 11673, 240, 20485, 35102, 51144 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 92, "seek": 31140, "start": 326.97999999999996, "end": 329.35999999999996, "text": "首先要在右上角", "tokens": [ 51144, 36490, 4275, 3581, 22060, 5708, 29389, 51263 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 93, "seek": 31140, "start": 329.35999999999996, "end": 330.56, "text": "启用开发者模式", "tokens": [ 51263, 2392, 107, 9254, 18937, 28926, 12444, 41908, 27584, 51323 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 94, "seek": 31140, "start": 330.56, "end": 332.4, "text": "然后点击load unpacked", "tokens": [ 51323, 26636, 12579, 6336, 119, 2907, 26699, 292, 51415 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 95, "seek": 31140, "start": 333.03999999999996, "end": 335.46, "text": "在open real-time translate目录当中", "tokens": [ 51447, 3581, 15752, 957, 12, 3766, 13799, 11386, 7391, 243, 16233, 5975, 51568 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 96, "seek": 31140, "start": 335.46, "end": 336.85999999999996, "text": "选择dist", "tokens": [ 51568, 2215, 231, 6852, 102, 42649, 51638 ], "temperature": 0, "avg_logprob": -0.1670561689671462, "compression_ratio": 1.0775193798449612, "no_speech_prob": 1.9840643017410287e-11 }, { "id": 97, "seek": 33686, "start": 336.86, "end": 338.90000000000003, "text": "这个目录呢", "tokens": [ 50365, 15368, 11386, 7391, 243, 6240, 50467 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 98, "seek": 33686, "start": 338.90000000000003, "end": 340.66, "text": "就是打包后生成的目录", "tokens": [ 50467, 5620, 12467, 23305, 13547, 8244, 11336, 1546, 11386, 7391, 243, 50555 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 99, "seek": 33686, "start": 340.66, "end": 342.88, "text": "要使用这个插件也很简单", "tokens": [ 50555, 4275, 22982, 9254, 15368, 11673, 240, 20485, 6404, 4563, 11249, 222, 47446, 50666 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 100, "seek": 33686, "start": 342.88, "end": 344.62, "text": "在扩展这里点击它", "tokens": [ 50666, 3581, 3416, 102, 43491, 35102, 12579, 6336, 119, 11284, 50753 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 101, "seek": 33686, "start": 344.62, "end": 346.2, "text": "第一次使用的时候呢", "tokens": [ 50753, 18049, 9487, 22982, 9254, 49873, 6240, 50832 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 102, "seek": 33686, "start": 346.2, "end": 347.16, "text": "需要配个api key", "tokens": [ 50832, 35748, 38846, 7549, 35891, 2141, 50880 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 103, "seek": 33686, "start": 347.16, "end": 350.24, "text": "大家去寻一个自己的open-air的api key", "tokens": [ 50880, 6868, 6734, 4510, 119, 20182, 17645, 1546, 15752, 12, 1246, 1546, 35891, 2141, 51034 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 104, "seek": 33686, "start": 350.24, "end": 351.16, "text": "配置进来保存", "tokens": [ 51034, 38846, 34719, 36700, 6912, 24302, 39781, 51080 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 105, "seek": 33686, "start": 351.16, "end": 354.58000000000004, "text": "然后任意打开一个标签页", "tokens": [ 51080, 26636, 26443, 9042, 12467, 18937, 20182, 162, 3921, 7973, 122, 10178, 113, 51251 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 106, "seek": 33686, "start": 354.58000000000004, "end": 357.3, "text": "比如播放一段YouTube视频", "tokens": [ 51251, 36757, 49993, 12744, 2257, 28427, 3223, 41173, 40656, 39752, 51387 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 107, "seek": 33686, "start": 357.3, "end": 359.12, "text": "那我们现在呢", "tokens": [ 51387, 4184, 15003, 25040, 6240, 51478 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 108, "seek": 33686, "start": 359.12, "end": 360.38, "text": "再来做一番演示", "tokens": [ 51478, 8623, 6912, 10907, 2257, 33118, 31382, 25696, 51541 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 109, "seek": 33686, "start": 360.38, "end": 363.72, "text": "比如我想将我这个语音呢", "tokens": [ 51541, 36757, 25246, 45456, 1654, 15368, 5233, 255, 18034, 6240, 51708 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 110, "seek": 33686, "start": 363.72, "end": 365.22, "text": "实时的转换成日语", "tokens": [ 51708, 24726, 15729, 1546, 17819, 105, 26075, 95, 11336, 6890, 5233, 255, 51783 ], "temperature": 0, "avg_logprob": -0.13324017138094515, "compression_ratio": 1.2203947368421053, "no_speech_prob": 1.850372632894448e-11 }, { "id": 111, "seek": 36522, "start": 365.22, "end": 368.22, "text": "来听听效果吧", "tokens": [ 50365, 6912, 31022, 31022, 43076, 9319, 6062, 50515 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 112, "seek": 36522, "start": 368.22, "end": 370.66, "text": "音質的には", "tokens": [ 50515, 18034, 43450, 1546, 34779, 50637 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 113, "seek": 36522, "start": 370.66, "end": 373.3, "text": "丹丸コーポレートアイデンティティは持たず", "tokens": [ 50637, 940, 117, 940, 116, 18066, 15266, 251, 16680, 38551, 12817, 8040, 31327, 14121, 228, 26760, 22985, 26760, 3065, 17694, 3368, 18216, 50769 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 114, "seek": 36522, "start": 373.3, "end": 376.56, "text": "既存のコーポレートアイデンティティ機能や能力を活用して", "tokens": [ 50769, 4479, 95, 39781, 2972, 18066, 15266, 251, 16680, 38551, 12817, 8040, 31327, 14121, 228, 26760, 22985, 26760, 17543, 8225, 7355, 8225, 13486, 5998, 25956, 9254, 8822, 50932 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 115, "seek": 36522, "start": 376.56, "end": 378.6, "text": "open source のエコシステムにある", "tokens": [ 50932, 15752, 4009, 21171, 26215, 18066, 11054, 9550, 22985, 32026, 4108, 24719, 51034 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 116, "seek": 36522, "start": 378.6, "end": 381.22, "text": "非常に豊かなスキルマッケージを統合し", "tokens": [ 51034, 14392, 4108, 19517, 232, 24939, 9550, 15535, 9405, 13258, 8276, 32639, 44143, 5998, 33725, 14245, 2849, 51165 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 117, "seek": 36522, "start": 381.22, "end": 382.32000000000005, "text": "それによって", "tokens": [ 51165, 13873, 4108, 5591, 6102, 51220 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 118, "seek": 36522, "start": 382.32000000000005, "end": 383.84000000000003, "text": "クラウドデザインと似た", "tokens": [ 51220, 10825, 11353, 20745, 11195, 31327, 824, 114, 8040, 4824, 3193, 7384, 120, 3368, 51296 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 119, "seek": 36522, "start": 383.84000000000003, "end": 385.22, "text": "あるいはそれ以上に", "tokens": [ 51296, 24719, 1764, 3065, 13873, 29497, 4108, 51365 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 120, "seek": 36522, "start": 385.22, "end": 388.36, "text": "機能統合の豊かなopen source 精神を構築しました", "tokens": [ 51365, 17543, 8225, 33725, 14245, 2972, 19517, 232, 24939, 15752, 4009, 220, 34910, 20432, 5998, 43362, 20878, 231, 44499, 51522 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 121, "seek": 36522, "start": 388.36, "end": 389.82000000000005, "text": "open design は", "tokens": [ 51522, 15752, 1715, 20785, 51595 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 122, "seek": 36522, "start": 389.82000000000005, "end": 392.08000000000004, "text": "現時点のopen source エコシステムで", "tokens": [ 51595, 9581, 6611, 12579, 2972, 15752, 4009, 220, 26215, 18066, 11054, 9550, 22985, 32026, 2474, 51708 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 123, "seek": 36522, "start": 392.08000000000004, "end": 393.02000000000004, "text": "非常に人気のある", "tokens": [ 51708, 14392, 4108, 4035, 25870, 2972, 24719, 51755 ], "temperature": 0, "avg_logprob": -0.13659942150115967, "compression_ratio": 1.6071428571428572, "no_speech_prob": 3.1171512593353157e-10 }, { "id": 124, "seek": 39302, "start": 393.02, "end": 395.02, "text": "いくつかのプロジェクトを統合しています", "tokens": [ 50365, 49394, 9335, 3703, 2972, 20953, 17164, 14655, 28571, 10825, 7588, 5998, 33725, 14245, 8822, 11267, 50465 ], "temperature": 0, "avg_logprob": -0.20700454711914062, "compression_ratio": 1.0357142857142858, "no_speech_prob": 2.2730617388333485e-10 }, { "id": 125, "seek": 39302, "start": 395.02, "end": 396.02, "text": "例えば", "tokens": [ 50465, 46629, 50515 ], "temperature": 0, "avg_logprob": -0.20700454711914062, "compression_ratio": 1.0357142857142858, "no_speech_prob": 2.2730617388333485e-10 }, { "id": 126, "seek": 39302, "start": 396.02, "end": 397.02, "text": "花玉デザインや", "tokens": [ 50515, 20127, 8051, 231, 31327, 824, 114, 8040, 4824, 7355, 50565 ], "temperature": 0, "avg_logprob": -0.20700454711914062, "compression_ratio": 1.0357142857142858, "no_speech_prob": 2.2730617388333485e-10 }, { "id": 127, "seek": 39702, "start": 397.02, "end": 399.02, "text": "北角師匠のPTTスキルなど", "tokens": [ 50365, 26668, 29389, 16276, 9937, 254, 2972, 47, 28178, 9550, 15535, 9405, 3203, 7594, 50465 ], "temperature": 0, "avg_logprob": -0.26488676071166994, "compression_ratio": 0.95, "no_speech_prob": 3.5253033825455304e-10 }, { "id": 128, "seek": 39702, "start": 399.02, "end": 402.02, "text": "open course design なども含まれます", "tokens": [ 50465, 15752, 1164, 1715, 16647, 7594, 4801, 2392, 104, 2889, 4132, 5368, 50615 ], "temperature": 0, "avg_logprob": -0.26488676071166994, "compression_ratio": 0.95, "no_speech_prob": 3.5253033825455304e-10 }, { "id": 129, "seek": 40202, "start": 402.02, "end": 403.65999999999997, "text": "細心的朋友可能注意到了", "tokens": [ 50365, 41588, 7945, 1546, 19828, 16657, 40794, 21381, 50447 ], "temperature": 0, "avg_logprob": -0.21429141875236266, "compression_ratio": 1.32, "no_speech_prob": 4.333875203155202e-10 }, { "id": 130, "seek": 40202, "start": 403.65999999999997, "end": 405.4, "text": "剛才我將視頻暂停了", "tokens": [ 50447, 16940, 18888, 1654, 34168, 27333, 29940, 23826, 224, 36135, 2289, 50534 ], "temperature": 0, "avg_logprob": -0.21429141875236266, "compression_ratio": 1.32, "no_speech_prob": 4.333875203155202e-10 }, { "id": 131, "seek": 40202, "start": 405.4, "end": 408.15999999999997, "text": "但是語音輸出还在持續了一小段時間", "tokens": [ 50534, 11189, 31348, 18034, 12290, 116, 7781, 14852, 3581, 17694, 29583, 2289, 2257, 7322, 28427, 20788, 50672 ], "temperature": 0, "avg_logprob": -0.21429141875236266, "compression_ratio": 1.32, "no_speech_prob": 4.333875203155202e-10 }, { "id": 132, "seek": 40202, "start": 408.15999999999997, "end": 411.44, "text": "剛才我們在介绍模型的时候也聊到这一塊", "tokens": [ 50672, 16940, 18888, 5884, 3581, 30312, 10115, 235, 41908, 39823, 49873, 6404, 40096, 4511, 5562, 2257, 26268, 50836 ], "temperature": 0, "avg_logprob": -0.21429141875236266, "compression_ratio": 1.32, "no_speech_prob": 4.333875203155202e-10 }, { "id": 133, "seek": 40202, "start": 411.44, "end": 415.62, "text": "模型会收集一段必要的上下文再做翻譯", "tokens": [ 50836, 41908, 39823, 12949, 18681, 26020, 2257, 28427, 28531, 4275, 1546, 5708, 4438, 17174, 8623, 10907, 42716, 11823, 107, 51045 ], "temperature": 0, "avg_logprob": -0.21429141875236266, "compression_ratio": 1.32, "no_speech_prob": 4.333875203155202e-10 }, { "id": 134, "seek": 40202, "start": 415.62, "end": 419.03999999999996, "text": "因此在这个翻譯和語音輸出上", "tokens": [ 51045, 8698, 17947, 3581, 15368, 42716, 11823, 107, 12565, 31348, 18034, 12290, 116, 7781, 5708, 51216 ], "temperature": 0, "avg_logprob": -0.21429141875236266, "compression_ratio": 1.32, "no_speech_prob": 4.333875203155202e-10 }, { "id": 135, "seek": 40202, "start": 419.03999999999996, "end": 420.79999999999995, "text": "就会有一点小小的延遲", "tokens": [ 51216, 3111, 12949, 32241, 12579, 7322, 7322, 1546, 17527, 114, 3330, 110, 51304 ], "temperature": 0, "avg_logprob": -0.21429141875236266, "compression_ratio": 1.32, "no_speech_prob": 4.333875203155202e-10 }, { "id": 136, "seek": 40202, "start": 420.79999999999995, "end": 422.74, "text": "这也是它的特点吧", "tokens": [ 51304, 5562, 22021, 45224, 17682, 12579, 6062, 51401 ], "temperature": 0, "avg_logprob": -0.21429141875236266, "compression_ratio": 1.32, "no_speech_prob": 4.333875203155202e-10 }, { "id": 137, "seek": 40202, "start": 422.74, "end": 424.65999999999997, "text": "来看看開銷情况吧", "tokens": [ 51401, 6912, 22884, 8949, 42041, 115, 46514, 6062, 51497 ], "temperature": 0, "avg_logprob": -0.21429141875236266, "compression_ratio": 1.32, "no_speech_prob": 4.333875203155202e-10 }, { "id": 138, "seek": 42466, "start": 424.66, "end": 428.16, "text": "我大概前前後後在視頻录制前的準備到", "tokens": [ 50365, 1654, 32044, 8945, 8945, 5661, 5661, 3581, 27333, 29940, 7391, 243, 25491, 8945, 1546, 36241, 4511, 50540 ], "temperature": 0, "avg_logprob": -0.35353220036599486, "compression_ratio": 1.1733870967741935, "no_speech_prob": 1.895498341619728e-10 }, { "id": 139, "seek": 42466, "start": 428.16, "end": 430.34000000000003, "text": "视频中的实际分享大概", "tokens": [ 50540, 40656, 39752, 5975, 1546, 24726, 8842, 227, 30855, 32044, 50649 ], "temperature": 0, "avg_logprob": -0.35353220036599486, "compression_ratio": 1.1733870967741935, "no_speech_prob": 1.895498341619728e-10 }, { "id": 140, "seek": 42466, "start": 431.14000000000004, "end": 434.98, "text": "在語音的輸出上花了可能有半小时", "tokens": [ 50689, 3581, 31348, 18034, 1546, 12290, 116, 7781, 5708, 20127, 2289, 16657, 2412, 30018, 7322, 15729, 50881 ], "temperature": 0, "avg_logprob": -0.35353220036599486, "compression_ratio": 1.1733870967741935, "no_speech_prob": 1.895498341619728e-10 }, { "id": 141, "seek": 42466, "start": 435.78000000000003, "end": 439.1, "text": "那么目前的开销是0.9亿美元", "tokens": [ 50921, 35693, 39004, 1546, 18937, 23049, 222, 1541, 15, 13, 24, 1369, 123, 39433, 51087 ], "temperature": 0, "avg_logprob": -0.35353220036599486, "compression_ratio": 1.1733870967741935, "no_speech_prob": 1.895498341619728e-10 }, { "id": 142, "seek": 42466, "start": 439.78000000000003, "end": 442.88, "text": "那现在就对咱们今天分享做一番简单的总结", "tokens": [ 51121, 4184, 25040, 3111, 8713, 8975, 109, 9497, 12074, 30855, 10907, 2257, 33118, 11249, 222, 47446, 1546, 33440, 45641, 51276 ], "temperature": 0, "avg_logprob": -0.35353220036599486, "compression_ratio": 1.1733870967741935, "no_speech_prob": 1.895498341619728e-10 }, { "id": 143, "seek": 42466, "start": 443.08000000000004, "end": 443.78000000000003, "text": "首先呢", "tokens": [ 51286, 36490, 6240, 51321 ], "temperature": 0, "avg_logprob": -0.35353220036599486, "compression_ratio": 1.1733870967741935, "no_speech_prob": 1.895498341619728e-10 }, { "id": 144, "seek": 42466, "start": 443.78000000000003, "end": 445.34000000000003, "text": "这个chrome插件", "tokens": [ 51321, 15368, 339, 11505, 11673, 240, 20485, 51399 ], "temperature": 0, "avg_logprob": -0.35353220036599486, "compression_ratio": 1.1733870967741935, "no_speech_prob": 1.895498341619728e-10 }, { "id": 145, "seek": 42466, "start": 445.54, "end": 448.78000000000003, "text": "目前只是抓取的浏览器tab中的音频", "tokens": [ 51409, 39004, 36859, 43069, 29436, 1546, 11038, 237, 6758, 230, 34386, 83, 455, 5975, 1546, 18034, 39752, 51571 ], "temperature": 0, "avg_logprob": -0.35353220036599486, "compression_ratio": 1.1733870967741935, "no_speech_prob": 1.895498341619728e-10 }, { "id": 146, "seek": 44878, "start": 448.78, "end": 451.14, "text": "像zoom桌面客户端", "tokens": [ 50365, 12760, 4765, 298, 162, 5475, 8833, 32316, 1486, 115, 11957, 107, 50483 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 147, "seek": 44878, "start": 451.14, "end": 452.38, "text": "spotify桌面", "tokens": [ 50483, 4952, 310, 2505, 162, 5475, 8833, 50545 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 148, "seek": 44878, "start": 452.38, "end": 454.02, "text": "本地视频播放器等等呢", "tokens": [ 50545, 8802, 10928, 40656, 39752, 49993, 12744, 34386, 36000, 6240, 50627 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 149, "seek": 44878, "start": 454.02, "end": 454.88, "text": "是抓不到的", "tokens": [ 50627, 1541, 43069, 35728, 1546, 50670 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 150, "seek": 44878, "start": 454.88, "end": 458.11999999999995, "text": "不过同样的逻辑或者是代码呢", "tokens": [ 50670, 1960, 16866, 13089, 14496, 1546, 2215, 119, 9830, 239, 31148, 1541, 19105, 23230, 223, 6240, 50832 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 151, "seek": 44878, "start": 458.11999999999995, "end": 460.38, "text": "可以应用到不同的应用场景中", "tokens": [ 50832, 6723, 44297, 9254, 4511, 47123, 1546, 44297, 9254, 50255, 50218, 5975, 50945 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 152, "seek": 44878, "start": 460.38, "end": 463.96, "text": "比如可以覆盖到我们刚才介绍的这些桌面应用", "tokens": [ 50945, 36757, 6723, 2888, 228, 5419, 244, 4511, 15003, 49160, 18888, 30312, 10115, 235, 1546, 5562, 13824, 162, 5475, 8833, 44297, 9254, 51124 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 153, "seek": 44878, "start": 463.96, "end": 467.91999999999996, "text": "至于具体的技术瓶颈在哪里", "tokens": [ 51124, 20844, 37732, 39806, 29485, 1546, 32502, 1474, 107, 163, 241, 114, 12501, 230, 3581, 17028, 15759, 51322 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 154, "seek": 44878, "start": 467.91999999999996, "end": 470.91999999999996, "text": "是否操作系统支持音频的捕获呢", "tokens": [ 51322, 1541, 45334, 13167, 235, 11914, 25368, 10115, 253, 35488, 18034, 39752, 1546, 26075, 243, 31127, 115, 6240, 51472 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 155, "seek": 44878, "start": 470.91999999999996, "end": 474.44, "text": "这个就是我们在技术或工程层面需要去考虑的了", "tokens": [ 51472, 15368, 5620, 15003, 3581, 32502, 1474, 107, 19780, 23323, 29649, 9636, 224, 8833, 35748, 6734, 26504, 12026, 239, 1546, 2289, 51648 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 156, "seek": 44878, "start": 474.44, "end": 477.41999999999996, "text": "第二点是GPT Real-Time Translate", "tokens": [ 51648, 19693, 12579, 1541, 38, 47, 51, 8467, 12, 22233, 6531, 17593, 51797 ], "temperature": 0, "avg_logprob": -0.14314335909756748, "compression_ratio": 1.2160493827160495, "no_speech_prob": 1.3066900859948216e-11 }, { "id": 157, "seek": 47742, "start": 477.42, "end": 479.02000000000004, "text": "这个实时翻译模型", "tokens": [ 50365, 15368, 24726, 15729, 42716, 5233, 239, 41908, 39823, 50445 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 158, "seek": 47742, "start": 479.02000000000004, "end": 481.5, "text": "它不支持自定义的提示词", "tokens": [ 50445, 11284, 1960, 35488, 9722, 12088, 2930, 231, 1546, 20949, 25696, 5233, 235, 50569 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 159, "seek": 47742, "start": 481.5, "end": 482.52000000000004, "text": "数语表", "tokens": [ 50569, 33188, 5233, 255, 17571, 50620 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 160, "seek": 47742, "start": 482.52000000000004, "end": 483.62, "text": "配音选择", "tokens": [ 50620, 38846, 18034, 2215, 231, 6852, 102, 50675 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 161, "seek": 47742, "start": 483.62, "end": 486.62, "text": "如果你的场景对专有名词敏感", "tokens": [ 50675, 13119, 18961, 50255, 50218, 8713, 940, 241, 2412, 15940, 5233, 235, 7017, 237, 9709, 50825 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 162, "seek": 47742, "start": 486.62, "end": 491.18, "text": "它会按一般的翻译习惯来处理", "tokens": [ 50825, 11284, 12949, 26613, 2257, 49640, 1546, 42716, 5233, 239, 2930, 254, 4799, 107, 6912, 1787, 226, 13876, 51053 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 163, "seek": 47742, "start": 491.18, "end": 493.62, "text": "另外一点很重要的是", "tokens": [ 51053, 26202, 49042, 4563, 24928, 24620, 51175 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 164, "seek": 47742, "start": 493.62, "end": 495.3, "text": "Session没有关掉", "tokens": [ 51175, 50, 4311, 17944, 28053, 29327, 51259 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 165, "seek": 47742, "start": 495.3, "end": 496.04, "text": "它会一直跑", "tokens": [ 51259, 11284, 12949, 34448, 32585, 51296 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 166, "seek": 47742, "start": 496.04, "end": 496.8, "text": "一直击费", "tokens": [ 51296, 34448, 6336, 119, 18464, 117, 51334 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 167, "seek": 47742, "start": 496.8, "end": 500.12, "text": "一小时大概2美元", "tokens": [ 51334, 2257, 7322, 15729, 32044, 17, 39433, 51500 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 168, "seek": 47742, "start": 500.12, "end": 502.06, "text": "四小时大概就是8美元", "tokens": [ 51500, 19425, 7322, 15729, 32044, 5620, 23, 39433, 51597 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 169, "seek": 47742, "start": 502.06, "end": 503.04, "text": "因此呢", "tokens": [ 51597, 8698, 17947, 6240, 51646 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 170, "seek": 47742, "start": 503.04, "end": 504.5, "text": "大家用完记得要Stop", "tokens": [ 51646, 6868, 9254, 14128, 34756, 5916, 4275, 26163, 51719 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 171, "seek": 47742, "start": 504.5, "end": 507.06, "text": "在插件层面", "tokens": [ 51719, 3581, 11673, 240, 20485, 9636, 224, 8833, 51847 ], "temperature": 0, "avg_logprob": -0.12340267611221528, "compression_ratio": 1.2103321033210332, "no_speech_prob": 1.8063668616452588e-11 }, { "id": 172, "seek": 50706, "start": 507.06, "end": 509.5, "text": "大家就记得点击Stop Translating", "tokens": [ 50365, 6868, 3111, 34756, 5916, 12579, 6336, 119, 26163, 6531, 75, 990, 50487 ], "temperature": 0, "avg_logprob": -0.24890340760696766, "compression_ratio": 1.0414507772020725, "no_speech_prob": 1.5078180023997412e-11 }, { "id": 173, "seek": 50706, "start": 509.5, "end": 510.38, "text": "Restop", "tokens": [ 50487, 49, 377, 404, 50531 ], "temperature": 0, "avg_logprob": -0.24890340760696766, "compression_ratio": 1.0414507772020725, "no_speech_prob": 1.5078180023997412e-11 }, { "id": 174, "seek": 50706, "start": 510.38, "end": 512.54, "text": "如果大家不再使用了呢", "tokens": [ 50531, 13119, 6868, 1960, 8623, 22982, 9254, 2289, 6240, 50639 ], "temperature": 0, "avg_logprob": -0.24890340760696766, "compression_ratio": 1.0414507772020725, "no_speech_prob": 1.5078180023997412e-11 }, { "id": 175, "seek": 50706, "start": 512.54, "end": 513.56, "text": "点击Stop", "tokens": [ 50639, 12579, 6336, 119, 26163, 50690 ], "temperature": 0, "avg_logprob": -0.24890340760696766, "compression_ratio": 1.0414507772020725, "no_speech_prob": 1.5078180023997412e-11 }, { "id": 176, "seek": 50706, "start": 513.56, "end": 514.22, "text": "停掉它", "tokens": [ 50690, 36135, 29327, 11284, 50723 ], "temperature": 0, "avg_logprob": -0.24890340760696766, "compression_ratio": 1.0414507772020725, "no_speech_prob": 1.5078180023997412e-11 }, { "id": 177, "seek": 50706, "start": 514.22, "end": 515.08, "text": "否则呢", "tokens": [ 50723, 45334, 2437, 247, 6240, 50766 ], "temperature": 0, "avg_logprob": -0.24890340760696766, "compression_ratio": 1.0414507772020725, "no_speech_prob": 1.5078180023997412e-11 }, { "id": 178, "seek": 50706, "start": 515.08, "end": 516.7, "text": "这个Session一直会开着", "tokens": [ 50766, 15368, 50, 4311, 34448, 12949, 18937, 20708, 50847 ], "temperature": 0, "avg_logprob": -0.24890340760696766, "compression_ratio": 1.0414507772020725, "no_speech_prob": 1.5078180023997412e-11 }, { "id": 179, "seek": 50706, "start": 516.7, "end": 519.46, "text": "整个扩展从想法到能用", "tokens": [ 50847, 27662, 7549, 3416, 102, 43491, 35630, 7093, 11148, 4511, 8225, 9254, 50985 ], "temperature": 0, "avg_logprob": -0.24890340760696766, "compression_ratio": 1.0414507772020725, "no_speech_prob": 1.5078180023997412e-11 }, { "id": 180, "seek": 50706, "start": 519.46, "end": 521.96, "text": "目前在AEA时代编写起来非常的快", "tokens": [ 50985, 39004, 3581, 32, 36, 32, 15729, 19105, 38109, 244, 5676, 247, 9147, 6912, 48263, 10251, 51110 ], "temperature": 0, "avg_logprob": -0.24890340760696766, "compression_ratio": 1.0414507772020725, "no_speech_prob": 1.5078180023997412e-11 }, { "id": 181, "seek": 52196, "start": 521.96, "end": 523.96, "text": "给到它足够的参考文档", "tokens": [ 50365, 23197, 4511, 11284, 37236, 1787, 253, 1546, 2129, 224, 26504, 17174, 21416, 96, 50465 ], "temperature": 0, "avg_logprob": -0.1639658959357293, "compression_ratio": 1.0427807486631016, "no_speech_prob": 1.751212716505357e-11 }, { "id": 182, "seek": 52196, "start": 523.96, "end": 524.86, "text": "API手册", "tokens": [ 50465, 4715, 40, 11389, 5676, 234, 50510 ], "temperature": 0, "avg_logprob": -0.1639658959357293, "compression_ratio": 1.0427807486631016, "no_speech_prob": 1.751212716505357e-11 }, { "id": 183, "seek": 52196, "start": 524.86, "end": 526.2, "text": "很快就搞定了", "tokens": [ 50510, 4563, 10251, 3111, 41816, 12088, 2289, 50577 ], "temperature": 0, "avg_logprob": -0.1639658959357293, "compression_ratio": 1.0427807486631016, "no_speech_prob": 1.751212716505357e-11 }, { "id": 184, "seek": 52196, "start": 526.2, "end": 528.48, "text": "你可以把它分叉下来", "tokens": [ 50577, 42766, 42061, 6627, 2129, 231, 4438, 6912, 50691 ], "temperature": 0, "avg_logprob": -0.1639658959357293, "compression_ratio": 1.0427807486631016, "no_speech_prob": 1.751212716505357e-11 }, { "id": 185, "seek": 52196, "start": 528.48, "end": 530.6, "text": "加自己想要的功能", "tokens": [ 50691, 9990, 17645, 7093, 4275, 1546, 32311, 8225, 50797 ], "temperature": 0, "avg_logprob": -0.1639658959357293, "compression_ratio": 1.0427807486631016, "no_speech_prob": 1.751212716505357e-11 }, { "id": 186, "seek": 52196, "start": 530.6, "end": 532.64, "text": "比如把字幕导出到SRT", "tokens": [ 50797, 36757, 16075, 22381, 40211, 4510, 120, 7781, 4511, 50, 49, 51, 50899 ], "temperature": 0, "avg_logprob": -0.1639658959357293, "compression_ratio": 1.0427807486631016, "no_speech_prob": 1.751212716505357e-11 }, { "id": 187, "seek": 52196, "start": 532.64, "end": 534.62, "text": "加Hockey切换语言", "tokens": [ 50899, 9990, 39, 46164, 23632, 26075, 95, 5233, 255, 12009, 50998 ], "temperature": 0, "avg_logprob": -0.1639658959357293, "compression_ratio": 1.0427807486631016, "no_speech_prob": 1.751212716505357e-11 }, { "id": 188, "seek": 52196, "start": 534.62, "end": 538.2800000000001, "text": "把意音导进OBS做直播同传等等", "tokens": [ 50998, 16075, 9042, 18034, 4510, 120, 36700, 46, 8176, 10907, 16186, 49993, 13089, 7384, 254, 36000, 51181 ], "temperature": 0, "avg_logprob": -0.1639658959357293, "compression_ratio": 1.0427807486631016, "no_speech_prob": 1.751212716505357e-11 }, { "id": 189, "seek": 53828, "start": 538.28, "end": 540.72, "text": "这都是直接的扩展点", "tokens": [ 50365, 5562, 22796, 43297, 1546, 3416, 102, 43491, 12579, 50487 ], "temperature": 0, "avg_logprob": -0.0858989879488945, "compression_ratio": 1.1929133858267718, "no_speech_prob": 1.8691593411945817e-11 }, { "id": 190, "seek": 53828, "start": 540.72, "end": 545.0799999999999, "text": "OpenAI把同传从需要专业人员的服务", "tokens": [ 50487, 45569, 48698, 16075, 13089, 7384, 254, 35630, 35748, 940, 241, 940, 248, 4035, 3606, 246, 1546, 27408, 5087, 94, 50705 ], "temperature": 0, "avg_logprob": -0.0858989879488945, "compression_ratio": 1.1929133858267718, "no_speech_prob": 1.8691593411945817e-11 }, { "id": 191, "seek": 53828, "start": 545.0799999999999, "end": 546.54, "text": "变成了一个API调用", "tokens": [ 50705, 2129, 246, 11336, 2289, 20182, 4715, 40, 8897, 225, 9254, 50778 ], "temperature": 0, "avg_logprob": -0.0858989879488945, "compression_ratio": 1.1929133858267718, "no_speech_prob": 1.8691593411945817e-11 }, { "id": 192, "seek": 53828, "start": 546.54, "end": 548.88, "text": "这会不会是同传这个职业", "tokens": [ 50778, 5562, 12949, 47928, 1541, 13089, 7384, 254, 15368, 8171, 234, 940, 248, 50895 ], "temperature": 0, "avg_logprob": -0.0858989879488945, "compression_ratio": 1.1929133858267718, "no_speech_prob": 1.8691593411945817e-11 }, { "id": 193, "seek": 53828, "start": 548.88, "end": 553.16, "text": "也是跨语言内容消费方式的一个明显的转折点呢", "tokens": [ 50895, 22021, 6563, 101, 5233, 255, 12009, 34742, 25750, 28837, 18464, 117, 9249, 27584, 1546, 20182, 11100, 1431, 122, 1546, 17819, 105, 7235, 246, 12579, 6240, 51109 ], "temperature": 0, "avg_logprob": -0.0858989879488945, "compression_ratio": 1.1929133858267718, "no_speech_prob": 1.8691593411945817e-11 }, { "id": 194, "seek": 53828, "start": 553.16, "end": 554.72, "text": "大家怎么看呢", "tokens": [ 51109, 6868, 15282, 4200, 6240, 51187 ], "temperature": 0, "avg_logprob": -0.0858989879488945, "compression_ratio": 1.1929133858267718, "no_speech_prob": 1.8691593411945817e-11 }, { "id": 195, "seek": 53828, "start": 554.72, "end": 556.8, "text": "那感兴趣的朋友", "tokens": [ 51187, 4184, 9709, 2347, 112, 39835, 1546, 19828, 51291 ], "temperature": 0, "avg_logprob": -0.0858989879488945, "compression_ratio": 1.1929133858267718, "no_speech_prob": 1.8691593411945817e-11 }, { "id": 196, "seek": 53828, "start": 556.8, "end": 559.42, "text": "或者在这方面已经有应用经验的朋友", "tokens": [ 51291, 31148, 3581, 5562, 39604, 49161, 2412, 44297, 9254, 30276, 49657, 234, 1546, 19828, 51422 ], "temperature": 0, "avg_logprob": -0.0858989879488945, "compression_ratio": 1.1929133858267718, "no_speech_prob": 1.8691593411945817e-11 }, { "id": 197, "seek": 53828, "start": 559.42, "end": 561.04, "text": "欢迎在评论区留言", "tokens": [ 51422, 28566, 17699, 3581, 5233, 226, 7422, 118, 9937, 118, 43681, 51503 ], "temperature": 0, "avg_logprob": -0.0858989879488945, "compression_ratio": 1.1929133858267718, "no_speech_prob": 1.8691593411945817e-11 }, { "id": 198, "seek": 56104, "start": 561.04, "end": 562.8399999999999, "text": "那今天的分享就到这里", "tokens": [ 50365, 4184, 34947, 30855, 45918, 35102, 50455 ], "temperature": 0, "avg_logprob": -0.07028617858886718, "compression_ratio": 0.8571428571428571, "no_speech_prob": 1.3682169112960718e-11 }, { "id": 199, "seek": 56104, "start": 562.8399999999999, "end": 563.74, "text": "感谢大家收看", "tokens": [ 50455, 9709, 11340, 6868, 34333, 50500 ], "temperature": 0, "avg_logprob": -0.07028617858886718, "compression_ratio": 0.8571428571428571, "no_speech_prob": 1.3682169112960718e-11 }, { "id": 200, "seek": 56104, "start": 563.74, "end": 564.8199999999999, "text": "我们下期再见", "tokens": [ 50500, 15003, 4438, 16786, 44176, 50554 ], "temperature": 0, "avg_logprob": -0.07028617858886718, "compression_ratio": 0.8571428571428571, "no_speech_prob": 1.3682169112960718e-11 } ], "transcript_selection": { "schema": "videonote.transcript_selection.v1", "status": "SELECTED_WITH_WARNINGS", "selected_source": "whisper_provisional", "decision_reason": "published_with_warning_after_language_guard:requested_language_repair_inconclusive", "publication_allowed": true, "whisper_rejected": true, "publication_mode": "published_with_warnings" } }