{
  "text": "Hello 大家好我是周日金\nAI这两年离咱们普通用户越来越近\n现在甚至不需要单独买一台服务器\n我们就可以直接把手头闲置在Mac\n变成一台属于自己的本地AI服务器\n这期视频我会从零开始\n带大家把Olama、OpenWeb UI、REG、OCR、Whisper.cpp和N8N一步步串起来\n这套方案的重点就是\n模型和大部分的数据都运行在自己本地的Mac上\n也不要求特别高的配置\n如果大家对这类玩法感兴趣\n后面还可以继续带大家折腾本地图片和视频生成\n本地AI编程\n以及把AI介入Home Assistant\n通过自然语言控制家里的智能设备\n视频正式开始前\n先求一波一进三连加关注支持一下\n这样以后有新的教程更新\n也能第一时间受到推送\n非常感谢\n接下来我们就从我的这台Mac mini来开始\n在开始下载模型之前\n我们需要先在Mac上安装OLAMA\n你可以把它理解为\n让普通用户可以在本地运行AI大模型的运行环境\n直接前往OLAMA的官网下载对应的安装包\n然后拖动到MacOS用程式文件夹就可以了\n平时使用ChatGBT Cloud这类AI服务时\n模型都是运行在云端服务器上\n而OLAMA要做的事情\n就是把像Deepseek签问LAMA这类开源模型\n下载安装到你的Mac上\n让他们直接利用本机的CPU\n和GPU运行\n安装后Olama并打开之后\n下一步就是下载模型\n不过这里需要注意啊\n本DAI和ChatGPT这类云端的模型\n最大的区别\n就是模型大小会直接影响运行体验\n所以不能盲目追求越大的模型就越好\n我这里使用的是一台M4入门款的Mac mini\n采用16GB的同一列存\n加256GB的固态硬盘\n确认Olama的服务在正常运行后\n我们在MacOS的中段App中\n输入这三条命令\n它们分别负责不同的功能\n第一个是大约模型\n也就是本地AI的大脑\n负责日常聊天\n内容生成\n总业资料等任务\n其中4B代表40亿参数\nQ4KM代表经过4B特量化\n可以在降低模型这样的同时\n保持不错的运行效果\n是目前比较适合普通Mac的选择\n第二个是向量模型\n主要用于知识库检索\n如果你以后想让AI读取自己的PDF\n文档 视频稿件\n然后根据这些资料回答问题\n就需要通过embedding模型\n先帮AI找到相关内容\n第三个是欧塞AR文字试园模型\n负责处理图片\n扫描版PDF等内容\n把图片里的文字提许出来\n再交给语言模型进行分析\n这里之所以选择签问的AR模型\n是因为签问是目前比较成熟的\n开源大模型之一\n对中文理解和中文生成比较友好\n其次是它提供的模型规格丰富\n从4B到更大的参数版本\n可以适配不同配置的Mac\n还有一个是\n它在OLAMA里的支持比较完善\n安装和电用也都比较简单\n当然除了签问\nDeepSeq、Lama等模型也都可以运行\n大家可以根据自己的需求进行选择\n至于不同的Mac应该选择多大的模型\n主要看新变性能和统一内存的大小\n8GB内存的Mac建议先从4B级别模型开始\n16GB内存运行4B会相对轻松\n也可以在实际内存压力云雪情况下尝试9B\n而24GB、32GB以及更高内存的Mac\n则可以继续尝试更大的规格\n但具体体验还会受到上下文长度和后台任务的影响\n这里需要注意啊\n模型标签不能随便改数字\n如果继续使用千万3.5\n目前OLAMA还提供9B 27B 35B等更大规格\n具体名称一定以OLAMA模型库实际存在的标签为准\n那咱们还是先把本地AI服务器跑起来\n后续再根据实际需求升级模型\n接下来我们在Microsoft的中段APP中\n输入以下命令OLAMA LIST\n查看是否已经全部下载完成\n如果前面下载的模型都正常列出来\n就可以来简单试一下模型能不能在本地正常运行\n输入制传命令\nolama run\n千万3.5\n4b-q4km\n然后随便发一句话\n比如你好\n如果能正常给出回复就没问题了\n随后输入制传命令结束对话\nBye\n之后我们就可以停止运行\n以释放资源\n输入制传命令\nolama stop\n千万3.5\n4b-q4km\n然后再输入制传命令\nolama ps\n查看这个AR大模型是否仍在运行\n此时应该唯空\n那么就说明我们的大模型已经没有在运行了\n接下来咱们就要开始准备部署\n家庭局网AN对话见面了\n前面我们已经通过Olama把AI模型运行在Mac上\n但是目前还需要通过中端的进行操作\n对于日常使用来说不太方便\n所以接下来我们就要给他搭建一个\n类似ChatGB的网页界面\n这里使用OpenWebUI\n也就是让Olama负责运行AI模型\nOpenWebUI负责提供交互界面\n部署完成后\n家里的手机平板电脑都可以通过\n浏览器放完这台Mac mini\n直接使用自己的本地AI\n考虑到中国大陆用户的网络环境\n这里直接通过Python安装OpenWebUI\n同时Olama继续只监听本机地址\n不会直接报道网络中\n本教程固定使用OpenWebUI 0.11.0版本\n不使用latest或滚动更新版本\n避免版本变化导致操作流程不一炸\n首先准备Python环境\n如果你的Mac已经安装Homebrew\n则可以输入brew install python at 3.11\n如果没有安装Homebrew\n也可以直接前往Python官方网站\n下载安装Python 3.11对应的MacOS安装包\n后面如果会继续折腾这类工具\n建议提前装好Homebrew\n管理和升级会方便不少\n接下来创建OpenWeb UI独立运行环境\n避免影响系统里的其他python项目\n后续相关代码我不再一一进行解释\n如果有看不懂\n可以查看代码上方的文本注释\n然后通过轻滑镜像安装OpenWebUI\n安装完成后\n第一次启动将以只允许本机访问\n先完成账号和安全设置\n等命令跑起来以后\n不要关闭这个中端窗口\n因为OpenWeb UI的服务就运行在这里\n然后617打开127.0.0.1\n网好3000\n第一次进入需要充电账号\n第一个账号就会自动成为管理员\n进入管理后台后\n前往设置身份验证\n把newsign apps关闭\n也就是不允许新用户自由注册\n同时暂时不要安装来源不明的第三方Tools\nFunctions或Pipelines\n接着就可以在流量器打开127.0.0.1\n冒号11434\n-API-Version\n确认欧拉玛本地API正常响应\n回到OpenWeb UI\n在检查之前下载的前文3.5\n4B-Q4KM是否已经出现在母型列表当中\n确认没有问题后\n在中端App中\n按Ctrl加C停止OpenWeb UI\n接下来开放家庭局域网\n实际在系统设置网络中查看Mac mini的局域网IP\n实际操作需要替完成你自己的IP\n重新启动OpenWeb UI\n现在Mac mini本机可以继续访问127.0.0.1\n网号3000\n并使用刚才创建的管理员账号进行登录\n通过家庭网络里的手机\n电脑则可以通过Mac mini的IP地址加段稿号3000来访问\n同样使用刚才的账号来登录\n这时候可以随便问您问题\n看看局域网访问和模型交用是否正常\n如果下设备都无法打开\n依次可以清查几个地方\n最后提醒一下\n这里搭建的家庭局域网AI服务器\n不会把服务暴露到公网\n因此不需要进行路由器短口转发\n也不要直接把OpenWeb AI开放到互联网环境\n而这样配置后保持终端App不关闭\nAI模式运行在自己的Mac上\n数据只在家庭网络里进行流转\n同时手机平板和电脑都可以共享这台本地AI服务器\n前面OpenWeb UI已经可以这张访问了\n接下来我们再给整个家庭AI服务器\n建立一个统一的API接口\n这样以后不管是N8N手机会议指令\n还是其他支持OpenAI API的程序\n都统一连接OpenWeb UI\n不需要直接访问底层的奥拉玛\n那这样的好处是\n模型、账号和权限都可以在OpenWeb UI这一层统一管理\n后面更换模型时\n也不用重新修改每一个应用\n时间咱们回到OpenWeb UI\n点击左下角头像依次进入设置\n身份验证找到API密钥并打开\n如果界面提供API密钥端点限制\n也一并开启\n然后填写这个\n这样生成的key就只允许访问模型列表和聊天接口\n接下来进入账号API密钥\n点击右侧的显示创建新的API密钥\n生成密钥通常会以SK开头\n这个密钥相当于访问我们家庭i服务器的密码\n不要直接公开或者分享出去\n如果后面准备长期让N8N这类的自动化工具使用\n建议专门创建一个普通的本地账号\n再用这个账号生成API密钥\n不要长期使用管理员账号的密钥\n因为密钥会计权创建这个账号的权限\n而且一个账号通常只保留一个API密钥\n接着我们来测试接口是否正常\n打开一个新的终端APP窗口\n把下面的命令中的SK换成刚才生成的密钥\n如果结构正常\n终端会访问当前OpenWeb UI可以调用的模型列表\n这里可以顺便记一下返回结果里模型队的真实ID\n因为后面其他地方调用AI时就需要填写这个名称\n接下来再来测试聊天接口\n同样的记得把SK开头的这串换成刚才生成的密钥\n而model后面需要填写刚才模型列表里面实际反馈的ID\n如果你下载的是其他签问或者其他模型\n就替完成对应的模型\n如果最后反馈Json里能够看到API正常\n就说明这一套接口已经可以正常使用了\n签问3.5在开启思考时\n反馈内容里可能还会出现reasoning content这一类字段\n这是正常现象\n到这里我们就相当于给整个家庭AI服务器建立一个统一的入口\n最底层人是Olama负责运行模型\n上面的OpenWeb UI统一管理模型和API\n在网上则可以接N8N手机快递指令或者其他应用\n后面不管增加多少个客户端都尽量连接OpenWeb UI\n而不是让每台设备直接访问欧拉玛\n整个架构会更容易管理\n前面我们已经把本地AI和聊天的API都搭好了\n接下来就再补上一个非常熟能功能\nCell文档知识库\n也就是RAG\nRAG可以列成先从你的资料里找内容\n再交给AI声称回答\n这样以后不管是自己的PDF\n说明书工作文档还是视频稿\n都可以放进知识库\n让AI尽量根据这些资料来回答\n而不是只靠模型本身的记忆\n首先进入OpenWeb UI试试里的文档页面\n进入模型引擎选择OLAMA\nAPI密钥可以留空\n地址填写127.0.0.1\n冒号11434\n进入模型出入\n我们在视频一开始就下载好的\n轻文3-embadding0.6b\n而这里顺便解释一下\n这个模型不负责聊天\n它的工作是把文档找完成\n便于检索的向量数据\n只要你提问以后\n系统才能快速找到和问题最相关的资料\n其他参数可以先按照这一套来设置\n大部分是我认的\n这些参数主要是在检索效果和上下文本\n占用之间做一个比较稳妥的平衡\n比如快大小可以列成\n每次把文档切成多大片段\n快重叠则是在相邻片段之间\n保留的部分重复内容\n避免一句话刚好被重点切断\nTuboK代表每次提问\n最多取回多少个相关片段\n这里先设置为3\n并不是越大越好\n因为塞进太多资料\n反而会占模型有限的上下文空间\n接下来我们创建一个\n非常简单的测试文件\n方便确认RA器是否真的正常工作\n用任意文本编定器\n新建一个Text文本\n如果你用的是\nMicroOS自带的文本编辑App\n记得在菜单里选择制作成文本\n比如里面写上这些内容\n然后保存为家庭AI测试资料.txt\n回到OpenWebUI\n找到左侧边来的工作空间\n再点击顶部的知识库\n最后点击右上角的创建\n名称填写家庭AI测试库\n目标可以简单写真\n用于测试本地RIG文档检索和应用功能\n知识库创建完之后\n再点击右上角的加号\n选择上传文件\n把刚才的txt文件上传进去\n后面也可以上传带有文字层的PDF\nDocx或Markdown文档\n上乘完之后\n在新兑化的聊天输入框里\n通过警号手动附加这个支持库\n而我这里为了教程演示方便\n先手动选择\n因为小参数模型主动判断\n什么时候动用支持库\n有时候不够稳定\n然后我们可以问一个问题\n只依据支持库回答\n项目代号和设备维护时间\n氛围是什么\n请给出引用\n正常情况下\n耐心等待一会儿\n他应该回答项目代号是清零\n设备维护时间是每月第一\n周六上午十点\n而且回答旁边能够看到对应的资料引用\n那么这套最基础的REG知识库就算大奖成功了\n不过这样配置下来\n你可能会发现还有一个地方容易踩坑\n如果你的PDF是扫描件\n本身没有可以复制的文字层\n就不要拿它来测试REG了\n因为知识库首先需要能够读取里面的文字\n才能进行后续的处理\n所以这种文件就需要先经过OCR\n把图片里的文字设备成TXC或者Markdown\n再上传到知识库来进行处理\n接下来我们就继续把本地OCR的功能\n也接近来\n前面我们已经让AI能够读取次的文档\n接下来再补上图片理解和OCR\n也就是让这台本地AI不只能读文字\n还可以看图片\n那这里先区分两个概念\n普通图片理解更适合看一张图里有什么\n界面再显示什么\n再顺便识别其中的文字\nOCR则更专注于把图片里的文字\n表格或者图表内容尽可能提取出来\n先在OpenWeb UI里选择我们前面使用的\n千万305 4B-KL4KM\n如果你的Mac配置更高\n也可以用9B的版本\n上传一张同时包含物体和文字图片\n然后输入\n先描述画面内容\n再逐项提取可见文字\n不确定的部分标注为不确定\n不需要自信补圈\n这里可以分别拿产品包装\n软件界面截图\n中文表格\n或者已经做好隐私透明的发票\n和数据照片来测试一下\n这样可以比较直观的判断出\n模型到底是理解了图片内容\n还是只是把里面的文字给识别出来\n如果你出现了类似我这样的提示\n就前往OpenWeb UI的设置\n模型\n找到刚才你前面使用的模型\n编辑界面\n找到高级参数下的NUMCTX\nOlama\n针对我16GB的Mac mini\n将数字更改为8192\n然后重复上的操作就可以了\n如果我们的目标不是理解图片\n而是尽可能把图片里的文字给提取出来\n就可以使用我们前面已经下载好的GLM OCR\nOCR模型运行前\n如果较大的千万模型还占着同一印存\n先可以把它停掉\n接下来我们准备好需要识别图片\n放到一个统一的文件夹中\n比如我这个目录像\n对于我们Mac仿达中的用户文件夹\nAI test\n这里手动新建一个test的文件夹就可以了\n对于需要识别不同的内容图片\n比如普通收据表格图表\n都可以使用不同的命令\n注意命令结尾的文件名\n需要换成你自己图片的文件名\n第一次测试时\n建议每次只处理一张图片\n先确认整套工作流能够稳定运行\n如果遇到扫描PDF\n也不用一开始就折腾完整的PDF解析工具\n最简单办法就是\n先用Microx自带的预览打开PDF\n把其中一页导成PN器\n再交给GLM OCR进行处理\n等确认单页OCR没问题后\n再考虑批量处理整份文档\n识别出来的内容还可以继续利用\n保存成TXT或者Markdown\n再上传到我们上一节创建的知识库\n然后交给签问做摘要\n自端整理或者检索问答\n那么整个流程就传起来了\n如果OCR出现空白重复或者乱码\n可以先更新Olam\n然后每次只处理一张图片\n图片过大时\n可以把长辨锁到大约1600至2200像素\n并优先使用PNG或者GPG\n同时检查原图有没有明显模糊\n倾斜反光\n这些都会直接影响识别效果\n除了文字和图片之外\n本地还有一个非常实用的场景\n就是处理视频和录音\n比如把会议录音转成文字\n把视频自动生成字幕\n或者把自己的素材库快速整理成可搜索的文本\n这里我们使用Whisper.cpp\n它是Whisper模型C++实现的\n针对Apple Silicon提供Mental加速\n可以直接利用Mac的GPU进行本地转写\n首先安装需要用到的工具\nFMMP1G负责处理音频格式\nWhisper.cpp负责语音识别\n如果你的Mac已经安装Homebrew\n可以直接输入代码来进行安装\n如果你仍然没有安装Homebrew\n安装whisper.cpp可以走圆码编译路线\n不过对于普通用户来说\n这一步会比Homebrew更复杂一些\n还需要提前安装Xcode的Command Align Tools\n这里就不展开了\n安装完成后\n可以通过下面两个命令来确认是否正常\n接下来下载Whisper模型\n我这里选择的是这个\n这是一个支持中文和多语言的模型\n相比完整Large模型体积更小\n同时保留了比较好的转眼效果\n更适合Mac本地运行\n我们先创建存放模型和临时文件的目录\n然后下载模型\n下载完整后再计算一次SHA-1的值\n方便和whisper.cpp模型文档公开的接应值进行核对\n模型准备好后就可以开始处理视频\n假设你的视频文件位于这个文件夹\n也就是MicroOS用户目录的影片文件夹当中\n大家根据自己的实际文件存储位置进行修改\n首先使用FMMP1G把视频里的音频转换成Whisper更适合处理的格式\n这几个参数简单介绍一下\nVM表示去掉视频\n只保留音频\nAR16表示转换成16000赫兹的采样率\nAC1表示单身道\nPCM S16LE则是转换成Whisper\n更容易处理的WAB格式\n接下来运行Whisper\n同时生成文字稿和字幕文件\n-M用来指定我们前面下载的Whisper模型\n-F指定需要转写的音频文件\n-LZH告诉我们模型\n这段文字主要内容是中文\n-Prompt则可以提前提供一些\n容易识别错误的中有名词\n比如Mac mini\nApple Silicon\nOlama和千问\n对于科技类视频有其有用\n比如这里我就加上了我的Lookup Weekly\n后面的-otxt-oosrt和-ovtt\n分别让它同时生成TXT文字稿\nSRT字幕和VTT字幕\n最后通过-off\n指定输出文件的位置和名称\n运行完成后会生成这些文件\n其中TXT是纯文字稿\n适合后续交给AI总结或者整理\nSRT是常见的字幕格式\n可以直接导入紧接软件\nVTT是主要用于网页播放器的字幕\n如果是8GB的Mac\n也可以先尝试这里的Large-V3-Turbo-Q50电话模型\n如果实际转写时内存压力较大或者速度不理想\n再换成体积更小的Small多语言模型\n不要下载来带EAN后置的英文专用模型\n因为它只针对英文优化\n如果需要中文转写\n也不要开启翻译参数\n否则模型会把内容翻译成英文\n而不是保留原始的语言\n到这里整套本地i工作流基本已经闭环\n视频和录音可以通过Whisper\n转写文字\n再交给前文做总结\n整理\n最后也可以存入前面搭建的知识库\n实现自己本地的内容管理系统\n前面我们已经分别完成了本地云转写OpenWebUIAPI和签问调用\n至于使用N8N搭建一条自动化工作流\n上传音频或者视频后\nN8N自动调用Whisper来转写\n再把转写稿交给签问来整理\n最后返回摘要提纲标题和证捷建议\n不需要每次手动复制粘贴\n首先把前面安装好的Whisper.cpp\n变成一个可以被其他程序调用的本地API\n先尝试允许Mac本机访问\n保持这个中端窗口开启\n再新开一个中南app窗口进行测试\n如果能够正常返回转写结果\nWhisper API就已经运行成功\n后面N8N会运行在Docker容器里\n如果出现无法连接Whisper的情况\n可以按Ctrl加C停止服务\n再这样改\n这里需要特别注意\nWhisper server本身没有用户登录和全线验证的功能\n所以使用0.0.0.0时\n只建议放在可信的家庭局网内\n不要进行路由器端口转发\n也不要把8178端口直接暴露到公网\n接下来安装N8N\n这里使用Docker Desktop来运行\n并固定N8N版本\n避免后续更新导致教程界面或者节点配置发生变化\n先从Docker官网下载安装Docker Desktop\n偷动到应用程序无电夹并启动\n确认Docker正常运行后\n再打开终端App输入下面的命令\n接下来生成一段随机的加密密钥\n把生成的结果替换到下面的XXX\n然后我们访问127.0.0.1\n号5678\n第一次进入需要创建N8N的本地管理员账号\n接下来我们来新建一条workflow工作流\n这套流程一共只有四个节点\nWebhook\nwebhook HTTP request with perjuanxie\nHTTP request\nHTTP Request的签文整理\nRespond to Webhook\n第一个是Webhook节点\n它相当于整条工作流的入口\n以后只要把视频post到这个地址\n工作流就会自动开始\nWebhook的具体设置如下\n没提到的设置保持不便即可\n第二个节点添加http request\n名称可以改成whisper转写\n它负责把刚刚上传的音视频文件交给Mac上的whisper server\n第二个节点http request的具体设置如下\n没提到的设置保持不便即可\n值得注意的是n8nbinaryfile这类型中\ninput data field name需要填写webhook节点实际收到的二进制字段名称\n点击前一个webhook节点\n在output的binary中进行查看\n如果显示为data那就填写为data\n如果显示为Data0\n那就填写为Data0\n这个名称必须完全一致\n否则会出现\nThe item has no binary field的报错\n第三个节点继续添加http request\n名称改成前文整理\n这个节点负责把Whisper返回完整转写稿\n交给前面搭建好的OpenWeb UI API\n第三个节点http request具体设置如下\n没提到的设置保持不变即可\n接下来回到前面的JSON输入框\n把整个字段切换到expression模式\n再粘贴下面的表达式\n它的逻辑很简单\nN8N自动读取Whisper转写节点里面的text\n再连同预设提示词一起发给亲吻\n最后添加Respound to Webhook节点\n它负责把整条工作流结果返回给定用者\nRespound with选择Json\n然后把Respound body切换成Expression模式\n粘贴下面的表达式\nRespound code保持200即可\n工作流设置完成以后\n先不要急着正式启用\n在N8N里点击Listen for Test Event\n让Webhook等待一次测试请求\n然后回到终端执行下面的命令\n注意要把XAI Token后面的章位内容\n替换成你前面实际设置的随机自如串\n如果整个流程正常\n你最终应该会得到\n类似这样的结果\n测试成功以后回到N8N的工作流程页面\n点击右上角的Publish发布\n发布后就不需要每次先点击Lessens for Test Event了\n正式Webhook地址也会从Webhook Test AI Video Notes\n变成Webhook AI Video Notes\n以后只要把视频传到这个接口\n整套流程就会自动变成Whisper\n在本地把视频转成文字\n千万再在本地完成摘要\n提纲 标题和章节建议\n最后把转写稿和整理结果一起返回\n整个过程中的语音和文本内容\n都可以在自己的Mac上完成处理\n到这里整套工作流已经跑通了\n不过直接使用Curve调用正式外部后格式\n中端反围的是标准的JSON\n所以会看到大量的斜杠N\n引号和转译字符\n比如斜杠N其实就是换\n并不是烂码\n只是JSON保存Markdown和多项文字时的正常格式\n为让最终结果更加适合阅读\n我们可以再用JQ把JSON里的转写稿\n和签文之力结果提取出来\n如果已经安装Homebrew\n则可以执行Brew install GQ\n安装完整后\n把刚才电用正式工作流的命令稍微改一下\n如果只是想看签问整理后的结果\n也可以进一步讲话\n如果只需要Whisper的完整转眼稿\n也可以这么改\n哦对了如果你的Mac没有安装Homebrew\n也不一定非要额外安装JQ\n前面已经准备了Python 3.1\n可以直接用Python自带的JSON模块完成同样的处理\n视频最后在补一个实际使用时一定会碰到的问题\nMac mini重启以后\n哪些东西需要重新打开\n我们前面搭建的模型和数据都保存在硬盘里\n不需要重新安装或者重新下载\n但Olama,OpenWeb UI,Whisper Server和N8N这些运行中的服务需要重新启动\n建议按照下面这个顺序来\n首先启动最底层的Olama\n如果安装的是MicOS版的Olama\n直接打开Olama app即可\n也可以在中端里执行Open-AOlama\n然后可以检查Olama是否已经恢复\n如果能够正常返回Olama的版本信息\n就说明底层模型服务已经正常\n接下来再来启动OpenWeb UI\n充计后可以先确认MacMe的局域网IP地址有没有发生变化\n因为我们前面的启动命令里填写过这个地址\n如果地址没有发生变化\n就继续使用之前的IP\n如果变化了\n下面命令行里的也要一起替换\n这个中端装口需要保持运行\n然后617访问\n127.0.0.1网号3000\n局域网中的其他设备\n仍然访问MacMe的IP地址网号3000\n长期使用的话也可以在路由器里给Mac mini固定一个局网地址\n选择每次重建后来进行检查\n如果只是聊天知识库或者OpenWeb UI\n其实不用一直启动Whisper Server\n只有需要运行视频转写或者后面的N8n视频工作流时再打开即可\n同样这个中端创口需要保持运行\n可以打开另一个中端测试\n只要服务器能够这样响应\n就可以继续交给N8n来调用了\n接着打开Docker Desktop App\n因为我们创建N8n容器时已经设置了Restart Unless Stoped\n所以正常情况下\n只要Docker app启动\n之前的N8n容器也会自动恢复运行\n不需要重新执行那一串Docker run\n可以先检查Docker PS\n如果Liviar里能够看到N8n\n就说明N8n已经自动恢复\n如果没有启动\n可以手动执行Docker StartN8n\n然后留言器打开127.0.0.1-5678\n如果能够进入N8n\n就说明自动化服务已经恢复正常\n到这里我们就从一台普通的Mac mini开始\n完成一套相对完整的家庭本地AI服务\nOlama负责运行AI模型\nOpenWeb UI提供对话界面和统一API\nEmbedding和RAG负责自己的文档知识库\nGLM OCR处理图片和扫描件\nWhisper.cpp负责视频录音转写\n再通过N8N把这些能力串成一套简单的自动化工作流\n这套方案其实不只是在Mac上跑了一个AI模型\n当对话、文档、图片、映视频和自动化都连起来之后\n这台Mac mini才开始从一台跑模型的电脑\n变成一台可以长期放在家里\n由不同设备随时电用的AI服务器\n当然这只是一个非常技术的框架\n后面还可以继续加入更大的模型\n自己的资料库\n手机快据指令\n甚至更多自动化流程\n大家完全可以根据自己的Mac配置\n和实际需求继续往上搭\n先把这套技术环境稳定跑起来\n再逐步增加功能\n更适合第一次接触本地AI的各位\n好了\n那么以上就是本期视频的全部内容了\n如果觉得本期视频有所帮助的话\n还请不要忘记点赞链加工主支持一下\n就是李军\n我们下期视频再见\n拜拜",
  "language": "multilingual",
  "duration": 1666.751563,
  "segments": [
    {
      "start": 0.0,
      "end": 0.94,
      "text": "Hello 大家好我是周日金",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 0.94,
      "end": 3.62,
      "text": "AI这两年离咱们普通用户越来越近",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 3.62,
      "end": 6.08,
      "text": "现在甚至不需要单独买一台服务器",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 6.08,
      "end": 8.44,
      "text": "我们就可以直接把手头闲置在Mac",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 8.44,
      "end": 11.0,
      "text": "变成一台属于自己的本地AI服务器",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 11.0,
      "end": 13.0,
      "text": "这期视频我会从零开始",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 13.0,
      "end": 20.0,
      "text": "带大家把Olama、OpenWeb UI、REG、OCR、Whisper.cpp和N8N一步步串起来",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 20.0,
      "end": 21.42,
      "text": "这套方案的重点就是",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 21.42,
      "end": 25.42,
      "text": "模型和大部分的数据都运行在自己本地的Mac上",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 25.42,
      "end": 27.14,
      "text": "也不要求特别高的配置",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 27.14,
      "end": 29.16,
      "text": "如果大家对这类玩法感兴趣",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 29.16,
      "end": 32.78,
      "text": "后面还可以继续带大家折腾本地图片和视频生成",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 32.78,
      "end": 34.04,
      "text": "本地AI编程",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 34.04,
      "end": 36.18,
      "text": "以及把AI介入Home Assistant",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 36.18,
      "end": 38.8,
      "text": "通过自然语言控制家里的智能设备",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 38.8,
      "end": 40.02,
      "text": "视频正式开始前",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 40.02,
      "end": 42.18,
      "text": "先求一波一进三连加关注支持一下",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 42.18,
      "end": 43.92,
      "text": "这样以后有新的教程更新",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 43.92,
      "end": 45.34,
      "text": "也能第一时间受到推送",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 45.34,
      "end": 46.1,
      "text": "非常感谢",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 46.1,
      "end": 49.52,
      "text": "接下来我们就从我的这台Mac mini来开始",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 0,
      "language_run": 0,
      "source": "whisper"
    },
    {
      "start": 54.742,
      "end": 56.332,
      "text": "在开始下载模型之前",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 56.332,
      "end": 58.892,
      "text": "我们需要先在Mac上安装OLAMA",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 58.892,
      "end": 60.212,
      "text": "你可以把它理解为",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 60.212,
      "end": 63.992,
      "text": "让普通用户可以在本地运行AI大模型的运行环境",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 63.992,
      "end": 66.692,
      "text": "直接前往OLAMA的官网下载对应的安装包",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 66.692,
      "end": 70.252,
      "text": "然后拖动到MacOS用程式文件夹就可以了",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 70.252,
      "end": 72.952,
      "text": "平时使用ChatGBT Cloud这类AI服务时",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 72.952,
      "end": 75.272,
      "text": "模型都是运行在云端服务器上",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 75.272,
      "end": 76.952,
      "text": "而OLAMA要做的事情",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 76.952,
      "end": 80.772,
      "text": "就是把像Deepseek签问LAMA这类开源模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 80.772,
      "end": 82.212,
      "text": "下载安装到你的Mac上",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 82.212,
      "end": 84.252,
      "text": "让他们直接利用本机的CPU",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 84.252,
      "end": 85.332,
      "text": "和GPU运行",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 85.332,
      "end": 87.192,
      "text": "安装后Olama并打开之后",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 87.192,
      "end": 88.532,
      "text": "下一步就是下载模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 88.532,
      "end": 89.992,
      "text": "不过这里需要注意啊",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 89.992,
      "end": 92.572,
      "text": "本DAI和ChatGPT这类云端的模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 92.572,
      "end": 93.332,
      "text": "最大的区别",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 93.332,
      "end": 95.812,
      "text": "就是模型大小会直接影响运行体验",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 95.812,
      "end": 98.612,
      "text": "所以不能盲目追求越大的模型就越好",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 98.612,
      "end": 101.532,
      "text": "我这里使用的是一台M4入门款的Mac mini",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 101.532,
      "end": 103.072,
      "text": "采用16GB的同一列存",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 103.072,
      "end": 104.912,
      "text": "加256GB的固态硬盘",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 104.912,
      "end": 107.292,
      "text": "确认Olama的服务在正常运行后",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 107.292,
      "end": 109.132,
      "text": "我们在MacOS的中段App中",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 109.132,
      "end": 110.432,
      "text": "输入这三条命令",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 110.432,
      "end": 112.252,
      "text": "它们分别负责不同的功能",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 112.252,
      "end": 113.872,
      "text": "第一个是大约模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 113.872,
      "end": 115.692,
      "text": "也就是本地AI的大脑",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 115.692,
      "end": 116.772,
      "text": "负责日常聊天",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 116.772,
      "end": 117.492,
      "text": "内容生成",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 117.492,
      "end": 118.732,
      "text": "总业资料等任务",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 118.732,
      "end": 120.972,
      "text": "其中4B代表40亿参数",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 120.972,
      "end": 123.672,
      "text": "Q4KM代表经过4B特量化",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 123.672,
      "end": 125.632,
      "text": "可以在降低模型这样的同时",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 125.632,
      "end": 127.172,
      "text": "保持不错的运行效果",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 127.172,
      "end": 129.692,
      "text": "是目前比较适合普通Mac的选择",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 129.692,
      "end": 131.432,
      "text": "第二个是向量模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 131.432,
      "end": 132.832,
      "text": "主要用于知识库检索",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 132.832,
      "end": 135.312,
      "text": "如果你以后想让AI读取自己的PDF",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 135.312,
      "end": 136.592,
      "text": "文档 视频稿件",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 136.592,
      "end": 138.432,
      "text": "然后根据这些资料回答问题",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 138.432,
      "end": 140.072,
      "text": "就需要通过embedding模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 140.072,
      "end": 142.192,
      "text": "先帮AI找到相关内容",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 142.192,
      "end": 144.792,
      "text": "第三个是欧塞AR文字试园模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 144.792,
      "end": 145.912,
      "text": "负责处理图片",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 145.912,
      "end": 147.472,
      "text": "扫描版PDF等内容",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 147.472,
      "end": 149.332,
      "text": "把图片里的文字提许出来",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 149.332,
      "end": 151.332,
      "text": "再交给语言模型进行分析",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 151.332,
      "end": 153.612,
      "text": "这里之所以选择签问的AR模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 153.612,
      "end": 155.732,
      "text": "是因为签问是目前比较成熟的",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 155.732,
      "end": 156.752,
      "text": "开源大模型之一",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 156.752,
      "end": 159.332,
      "text": "对中文理解和中文生成比较友好",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 159.332,
      "end": 161.432,
      "text": "其次是它提供的模型规格丰富",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 161.432,
      "end": 163.712,
      "text": "从4B到更大的参数版本",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 163.712,
      "end": 165.372,
      "text": "可以适配不同配置的Mac",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 165.372,
      "end": 166.152,
      "text": "还有一个是",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 166.152,
      "end": 168.192,
      "text": "它在OLAMA里的支持比较完善",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 168.192,
      "end": 170.012,
      "text": "安装和电用也都比较简单",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 170.012,
      "end": 171.272,
      "text": "当然除了签问",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 171.272,
      "end": 173.812,
      "text": "DeepSeq、Lama等模型也都可以运行",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 173.812,
      "end": 176.092,
      "text": "大家可以根据自己的需求进行选择",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 176.092,
      "end": 179.012,
      "text": "至于不同的Mac应该选择多大的模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 179.012,
      "end": 181.252,
      "text": "主要看新变性能和统一内存的大小",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 181.252,
      "end": 184.452,
      "text": "8GB内存的Mac建议先从4B级别模型开始",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 184.452,
      "end": 187.872,
      "text": "16GB内存运行4B会相对轻松",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 187.872,
      "end": 191.332,
      "text": "也可以在实际内存压力云雪情况下尝试9B",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 191.332,
      "end": 194.632,
      "text": "而24GB、32GB以及更高内存的Mac",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 194.632,
      "end": 196.492,
      "text": "则可以继续尝试更大的规格",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 196.492,
      "end": 200.172,
      "text": "但具体体验还会受到上下文长度和后台任务的影响",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 200.172,
      "end": 201.532,
      "text": "这里需要注意啊",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 201.532,
      "end": 203.412,
      "text": "模型标签不能随便改数字",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 203.412,
      "end": 205.252,
      "text": "如果继续使用千万3.5",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 205.252,
      "end": 209.092,
      "text": "目前OLAMA还提供9B 27B 35B等更大规格",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 209.092,
      "end": 212.512,
      "text": "具体名称一定以OLAMA模型库实际存在的标签为准",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 212.512,
      "end": 215.152,
      "text": "那咱们还是先把本地AI服务器跑起来",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 215.152,
      "end": 217.412,
      "text": "后续再根据实际需求升级模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 217.412,
      "end": 219.692,
      "text": "接下来我们在Microsoft的中段APP中",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 219.692,
      "end": 221.872,
      "text": "输入以下命令OLAMA LIST",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 221.872,
      "end": 224.132,
      "text": "查看是否已经全部下载完成",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 224.132,
      "end": 226.632,
      "text": "如果前面下载的模型都正常列出来",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 226.632,
      "end": 229.892,
      "text": "就可以来简单试一下模型能不能在本地正常运行",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 229.892,
      "end": 231.072,
      "text": "输入制传命令",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 231.072,
      "end": 232.032,
      "text": "olama run",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 232.032,
      "end": 232.892,
      "text": "千万3.5",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 232.892,
      "end": 234.532,
      "text": "4b-q4km",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 234.532,
      "end": 235.892,
      "text": "然后随便发一句话",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 235.892,
      "end": 236.632,
      "text": "比如你好",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 236.632,
      "end": 239.072,
      "text": "如果能正常给出回复就没问题了",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 239.072,
      "end": 241.052,
      "text": "随后输入制传命令结束对话",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 241.052,
      "end": 241.532,
      "text": "Bye",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 241.532,
      "end": 243.652,
      "text": "之后我们就可以停止运行",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 243.652,
      "end": 244.632,
      "text": "以释放资源",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 244.632,
      "end": 245.772,
      "text": "输入制传命令",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 245.772,
      "end": 246.832,
      "text": "olama stop",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 246.832,
      "end": 247.652,
      "text": "千万3.5",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 247.652,
      "end": 249.072,
      "text": "4b-q4km",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 249.072,
      "end": 250.692,
      "text": "然后再输入制传命令",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 250.692,
      "end": 251.472,
      "text": "olama ps",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 251.472,
      "end": 254.132,
      "text": "查看这个AR大模型是否仍在运行",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 254.132,
      "end": 255.452,
      "text": "此时应该唯空",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 255.452,
      "end": 258.192,
      "text": "那么就说明我们的大模型已经没有在运行了",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 258.192,
      "end": 260.332,
      "text": "接下来咱们就要开始准备部署",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 260.332,
      "end": 262.332,
      "text": "家庭局网AN对话见面了",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 1,
      "language_run": 1,
      "source": "whisper"
    },
    {
      "start": 267.604,
      "end": 270.954,
      "text": "前面我们已经通过Olama把AI模型运行在Mac上",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 270.954,
      "end": 273.494,
      "text": "但是目前还需要通过中端的进行操作",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 273.494,
      "end": 275.274,
      "text": "对于日常使用来说不太方便",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 275.274,
      "end": 277.194,
      "text": "所以接下来我们就要给他搭建一个",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 277.194,
      "end": 279.054,
      "text": "类似ChatGB的网页界面",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 279.054,
      "end": 280.714,
      "text": "这里使用OpenWebUI",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 280.714,
      "end": 283.254,
      "text": "也就是让Olama负责运行AI模型",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 283.254,
      "end": 285.614,
      "text": "OpenWebUI负责提供交互界面",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 285.614,
      "end": 286.794,
      "text": "部署完成后",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 286.794,
      "end": 288.834,
      "text": "家里的手机平板电脑都可以通过",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 288.834,
      "end": 290.314,
      "text": "浏览器放完这台Mac mini",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 290.314,
      "end": 292.294,
      "text": "直接使用自己的本地AI",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 292.294,
      "end": 294.594,
      "text": "考虑到中国大陆用户的网络环境",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 294.594,
      "end": 296.914,
      "text": "这里直接通过Python安装OpenWebUI",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 296.914,
      "end": 299.434,
      "text": "同时Olama继续只监听本机地址",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 299.434,
      "end": 300.914,
      "text": "不会直接报道网络中",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 300.914,
      "end": 304.174,
      "text": "本教程固定使用OpenWebUI 0.11.0版本",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 304.174,
      "end": 306.574,
      "text": "不使用latest或滚动更新版本",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 306.574,
      "end": 309.394,
      "text": "避免版本变化导致操作流程不一炸",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 309.394,
      "end": 311.014,
      "text": "首先准备Python环境",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 311.014,
      "end": 313.014,
      "text": "如果你的Mac已经安装Homebrew",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 313.014,
      "end": 316.114,
      "text": "则可以输入brew install python at 3.11",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 316.114,
      "end": 317.554,
      "text": "如果没有安装Homebrew",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 317.554,
      "end": 319.494,
      "text": "也可以直接前往Python官方网站",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 319.494,
      "end": 322.174,
      "text": "下载安装Python 3.11对应的MacOS安装包",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 322.174,
      "end": 324.654,
      "text": "后面如果会继续折腾这类工具",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 324.654,
      "end": 326.114,
      "text": "建议提前装好Homebrew",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 326.114,
      "end": 327.854,
      "text": "管理和升级会方便不少",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 327.854,
      "end": 330.534,
      "text": "接下来创建OpenWeb UI独立运行环境",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 330.534,
      "end": 332.834,
      "text": "避免影响系统里的其他python项目",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 332.834,
      "end": 335.014,
      "text": "后续相关代码我不再一一进行解释",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 335.014,
      "end": 335.834,
      "text": "如果有看不懂",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 335.834,
      "end": 338.054,
      "text": "可以查看代码上方的文本注释",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 2,
      "language_run": 2,
      "source": "whisper"
    },
    {
      "start": 350.272,
      "end": 352.622,
      "text": "然后通过轻滑镜像安装OpenWebUI",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 3,
      "language_run": 3,
      "source": "whisper"
    },
    {
      "start": 362.916,
      "end": 363.806,
      "text": "安装完成后",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 4,
      "language_run": 4,
      "source": "whisper"
    },
    {
      "start": 363.806,
      "end": 366.026,
      "text": "第一次启动将以只允许本机访问",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 4,
      "language_run": 4,
      "source": "whisper"
    },
    {
      "start": 366.026,
      "end": 367.926,
      "text": "先完成账号和安全设置",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 4,
      "language_run": 4,
      "source": "whisper"
    },
    {
      "start": 367.926,
      "end": 369.306,
      "text": "等命令跑起来以后",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 4,
      "language_run": 4,
      "source": "whisper"
    },
    {
      "start": 369.306,
      "end": 370.886,
      "text": "不要关闭这个中端窗口",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 4,
      "language_run": 4,
      "source": "whisper"
    },
    {
      "start": 370.886,
      "end": 373.386,
      "text": "因为OpenWeb UI的服务就运行在这里",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 4,
      "language_run": 4,
      "source": "whisper"
    },
    {
      "start": 376.212,
      "end": 378.762,
      "text": "然后617打开127.0.0.1",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 5,
      "language_run": 5,
      "source": "whisper"
    },
    {
      "start": 378.762,
      "end": 379.582,
      "text": "网好3000",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 5,
      "language_run": 5,
      "source": "whisper"
    },
    {
      "start": 379.582,
      "end": 381.702,
      "text": "第一次进入需要充电账号",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 5,
      "language_run": 5,
      "source": "whisper"
    },
    {
      "start": 381.702,
      "end": 383.762,
      "text": "第一个账号就会自动成为管理员",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 5,
      "language_run": 5,
      "source": "whisper"
    },
    {
      "start": 387.421,
      "end": 388.451,
      "text": "进入管理后台后",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 388.451,
      "end": 389.931,
      "text": "前往设置身份验证",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 389.931,
      "end": 391.451,
      "text": "把newsign apps关闭",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 391.451,
      "end": 393.451,
      "text": "也就是不允许新用户自由注册",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 393.451,
      "end": 396.571,
      "text": "同时暂时不要安装来源不明的第三方Tools",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 396.571,
      "end": 397.991,
      "text": "Functions或Pipelines",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 397.991,
      "end": 400.991,
      "text": "接着就可以在流量器打开127.0.0.1",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 400.991,
      "end": 402.431,
      "text": "冒号11434",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 402.431,
      "end": 404.551,
      "text": "-API-Version",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 404.551,
      "end": 406.891,
      "text": "确认欧拉玛本地API正常响应",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 406.891,
      "end": 408.131,
      "text": "回到OpenWeb UI",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 408.131,
      "end": 410.371,
      "text": "在检查之前下载的前文3.5",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 410.371,
      "end": 413.771,
      "text": "4B-Q4KM是否已经出现在母型列表当中",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 413.771,
      "end": 414.891,
      "text": "确认没有问题后",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 414.891,
      "end": 415.871,
      "text": "在中端App中",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 415.871,
      "end": 418.411,
      "text": "按Ctrl加C停止OpenWeb UI",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 418.411,
      "end": 420.171,
      "text": "接下来开放家庭局域网",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 420.171,
      "end": 423.731,
      "text": "实际在系统设置网络中查看Mac mini的局域网IP",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 423.731,
      "end": 426.051,
      "text": "实际操作需要替完成你自己的IP",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 426.051,
      "end": 427.851,
      "text": "重新启动OpenWeb UI",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 6,
      "language_run": 6,
      "source": "whisper"
    },
    {
      "start": 431.0,
      "end": 434.35,
      "text": "现在Mac mini本机可以继续访问127.0.0.1",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 434.35,
      "end": 435.07,
      "text": "网号3000",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 435.07,
      "end": 437.95,
      "text": "并使用刚才创建的管理员账号进行登录",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 437.95,
      "end": 439.71,
      "text": "通过家庭网络里的手机",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 439.71,
      "end": 443.55,
      "text": "电脑则可以通过Mac mini的IP地址加段稿号3000来访问",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 443.55,
      "end": 445.51,
      "text": "同样使用刚才的账号来登录",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 445.51,
      "end": 446.91,
      "text": "这时候可以随便问您问题",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 446.91,
      "end": 449.37,
      "text": "看看局域网访问和模型交用是否正常",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 449.37,
      "end": 451.03,
      "text": "如果下设备都无法打开",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 451.03,
      "end": 452.35,
      "text": "依次可以清查几个地方",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 452.35,
      "end": 453.59,
      "text": "最后提醒一下",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 453.59,
      "end": 455.63,
      "text": "这里搭建的家庭局域网AI服务器",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 455.63,
      "end": 457.41,
      "text": "不会把服务暴露到公网",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 457.41,
      "end": 459.61,
      "text": "因此不需要进行路由器短口转发",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 459.61,
      "end": 462.53,
      "text": "也不要直接把OpenWeb AI开放到互联网环境",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 462.53,
      "end": 465.01,
      "text": "而这样配置后保持终端App不关闭",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 465.01,
      "end": 466.71,
      "text": "AI模式运行在自己的Mac上",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 466.71,
      "end": 468.95,
      "text": "数据只在家庭网络里进行流转",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 468.95,
      "end": 472.87,
      "text": "同时手机平板和电脑都可以共享这台本地AI服务器",
      "language": "zh",
      "chunk": 0,
      "local_language_run": 7,
      "language_run": 7,
      "source": "whisper"
    },
    {
      "start": 478.147,
      "end": 480.337,
      "text": "前面OpenWeb UI已经可以这张访问了",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 480.337,
      "end": 482.537,
      "text": "接下来我们再给整个家庭AI服务器",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 482.537,
      "end": 484.377,
      "text": "建立一个统一的API接口",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 484.377,
      "end": 486.937,
      "text": "这样以后不管是N8N手机会议指令",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 486.937,
      "end": 489.357,
      "text": "还是其他支持OpenAI API的程序",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 489.357,
      "end": 491.217,
      "text": "都统一连接OpenWeb UI",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 491.217,
      "end": 493.097,
      "text": "不需要直接访问底层的奥拉玛",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 493.097,
      "end": 494.117,
      "text": "那这样的好处是",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 494.117,
      "end": 497.797,
      "text": "模型、账号和权限都可以在OpenWeb UI这一层统一管理",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 497.797,
      "end": 499.037,
      "text": "后面更换模型时",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 499.037,
      "end": 501.097,
      "text": "也不用重新修改每一个应用",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 501.097,
      "end": 502.617,
      "text": "时间咱们回到OpenWeb UI",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 502.617,
      "end": 505.057,
      "text": "点击左下角头像依次进入设置",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 505.057,
      "end": 507.557,
      "text": "身份验证找到API密钥并打开",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 507.557,
      "end": 510.177,
      "text": "如果界面提供API密钥端点限制",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 510.177,
      "end": 511.137,
      "text": "也一并开启",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 511.137,
      "end": 512.477,
      "text": "然后填写这个",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 512.477,
      "end": 517.377,
      "text": "这样生成的key就只允许访问模型列表和聊天接口",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 517.377,
      "end": 519.317,
      "text": "接下来进入账号API密钥",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 519.317,
      "end": 522.037,
      "text": "点击右侧的显示创建新的API密钥",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 522.037,
      "end": 524.357,
      "text": "生成密钥通常会以SK开头",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 524.357,
      "end": 527.357,
      "text": "这个密钥相当于访问我们家庭i服务器的密码",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 527.357,
      "end": 529.317,
      "text": "不要直接公开或者分享出去",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 529.317,
      "end": 532.657,
      "text": "如果后面准备长期让N8N这类的自动化工具使用",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 532.657,
      "end": 535.157,
      "text": "建议专门创建一个普通的本地账号",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 535.157,
      "end": 537.197,
      "text": "再用这个账号生成API密钥",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 537.197,
      "end": 539.377,
      "text": "不要长期使用管理员账号的密钥",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 539.377,
      "end": 542.057,
      "text": "因为密钥会计权创建这个账号的权限",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 542.057,
      "end": 544.877,
      "text": "而且一个账号通常只保留一个API密钥",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 544.877,
      "end": 546.957,
      "text": "接着我们来测试接口是否正常",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 546.957,
      "end": 548.997,
      "text": "打开一个新的终端APP窗口",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 548.997,
      "end": 552.297,
      "text": "把下面的命令中的SK换成刚才生成的密钥",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 552.297,
      "end": 553.597,
      "text": "如果结构正常",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 553.597,
      "end": 556.897,
      "text": "终端会访问当前OpenWeb UI可以调用的模型列表",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 556.897,
      "end": 560.297,
      "text": "这里可以顺便记一下返回结果里模型队的真实ID",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 560.297,
      "end": 563.477,
      "text": "因为后面其他地方调用AI时就需要填写这个名称",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 563.477,
      "end": 565.517,
      "text": "接下来再来测试聊天接口",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 565.517,
      "end": 568.917,
      "text": "同样的记得把SK开头的这串换成刚才生成的密钥",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 568.917,
      "end": 572.477,
      "text": "而model后面需要填写刚才模型列表里面实际反馈的ID",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 572.477,
      "end": 575.057,
      "text": "如果你下载的是其他签问或者其他模型",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 575.057,
      "end": 576.377,
      "text": "就替完成对应的模型",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 576.377,
      "end": 579.117,
      "text": "如果最后反馈Json里能够看到API正常",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 579.117,
      "end": 581.757,
      "text": "就说明这一套接口已经可以正常使用了",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 581.757,
      "end": 583.797,
      "text": "签问3.5在开启思考时",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 583.797,
      "end": 586.997,
      "text": "反馈内容里可能还会出现reasoning content这一类字段",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 586.997,
      "end": 588.077,
      "text": "这是正常现象",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 588.077,
      "end": 592.357,
      "text": "到这里我们就相当于给整个家庭AI服务器建立一个统一的入口",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 592.357,
      "end": 594.897,
      "text": "最底层人是Olama负责运行模型",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 594.897,
      "end": 598.157,
      "text": "上面的OpenWeb UI统一管理模型和API",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 598.157,
      "end": 601.797,
      "text": "在网上则可以接N8N手机快递指令或者其他应用",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 601.797,
      "end": 605.277,
      "text": "后面不管增加多少个客户端都尽量连接OpenWeb UI",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 605.277,
      "end": 607.637,
      "text": "而不是让每台设备直接访问欧拉玛",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 607.637,
      "end": 609.517,
      "text": "整个架构会更容易管理",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 1,
      "language_run": 8,
      "source": "whisper"
    },
    {
      "start": 614.685,
      "end": 617.535,
      "text": "前面我们已经把本地AI和聊天的API都搭好了",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 617.535,
      "end": 619.975,
      "text": "接下来就再补上一个非常熟能功能",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 619.975,
      "end": 621.315,
      "text": "Cell文档知识库",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 621.315,
      "end": 622.515,
      "text": "也就是RAG",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 622.515,
      "end": 625.455,
      "text": "RAG可以列成先从你的资料里找内容",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 625.455,
      "end": 626.915,
      "text": "再交给AI声称回答",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 626.915,
      "end": 628.655,
      "text": "这样以后不管是自己的PDF",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 628.655,
      "end": 630.855,
      "text": "说明书工作文档还是视频稿",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 630.855,
      "end": 632.035,
      "text": "都可以放进知识库",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 632.035,
      "end": 634.255,
      "text": "让AI尽量根据这些资料来回答",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 634.255,
      "end": 636.255,
      "text": "而不是只靠模型本身的记忆",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 636.255,
      "end": 638.975,
      "text": "首先进入OpenWeb UI试试里的文档页面",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 638.975,
      "end": 640.875,
      "text": "进入模型引擎选择OLAMA",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 640.875,
      "end": 642.235,
      "text": "API密钥可以留空",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 642.235,
      "end": 644.675,
      "text": "地址填写127.0.0.1",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 644.675,
      "end": 646.035,
      "text": "冒号11434",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 646.035,
      "end": 647.095,
      "text": "进入模型出入",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 647.095,
      "end": 648.815,
      "text": "我们在视频一开始就下载好的",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 648.815,
      "end": 650.895,
      "text": "轻文3-embadding0.6b",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 650.895,
      "end": 652.455,
      "text": "而这里顺便解释一下",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 652.455,
      "end": 653.875,
      "text": "这个模型不负责聊天",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 653.875,
      "end": 655.735,
      "text": "它的工作是把文档找完成",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 655.735,
      "end": 657.635,
      "text": "便于检索的向量数据",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 657.635,
      "end": 658.495,
      "text": "只要你提问以后",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 658.495,
      "end": 661.135,
      "text": "系统才能快速找到和问题最相关的资料",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 661.135,
      "end": 663.275,
      "text": "其他参数可以先按照这一套来设置",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 663.275,
      "end": 664.635,
      "text": "大部分是我认的",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 664.635,
      "end": 667.195,
      "text": "这些参数主要是在检索效果和上下文本",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 667.195,
      "end": 669.215,
      "text": "占用之间做一个比较稳妥的平衡",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 669.215,
      "end": 670.755,
      "text": "比如快大小可以列成",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 670.755,
      "end": 672.675,
      "text": "每次把文档切成多大片段",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 672.675,
      "end": 674.815,
      "text": "快重叠则是在相邻片段之间",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 674.815,
      "end": 676.015,
      "text": "保留的部分重复内容",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 676.015,
      "end": 677.855,
      "text": "避免一句话刚好被重点切断",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 677.855,
      "end": 679.435,
      "text": "TuboK代表每次提问",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 679.435,
      "end": 681.515,
      "text": "最多取回多少个相关片段",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 681.515,
      "end": 682.975,
      "text": "这里先设置为3",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 682.975,
      "end": 684.195,
      "text": "并不是越大越好",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 684.195,
      "end": 685.575,
      "text": "因为塞进太多资料",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 685.575,
      "end": 687.975,
      "text": "反而会占模型有限的上下文空间",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 687.975,
      "end": 689.075,
      "text": "接下来我们创建一个",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 689.075,
      "end": 690.275,
      "text": "非常简单的测试文件",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 690.275,
      "end": 692.735,
      "text": "方便确认RA器是否真的正常工作",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 692.735,
      "end": 694.435,
      "text": "用任意文本编定器",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 694.435,
      "end": 695.855,
      "text": "新建一个Text文本",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 695.855,
      "end": 696.895,
      "text": "如果你用的是",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 696.895,
      "end": 698.695,
      "text": "MicroOS自带的文本编辑App",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 698.695,
      "end": 700.835,
      "text": "记得在菜单里选择制作成文本",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 700.835,
      "end": 702.575,
      "text": "比如里面写上这些内容",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 702.575,
      "end": 705.435,
      "text": "然后保存为家庭AI测试资料.txt",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 705.435,
      "end": 706.615,
      "text": "回到OpenWebUI",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 706.615,
      "end": 708.515,
      "text": "找到左侧边来的工作空间",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 708.515,
      "end": 710.075,
      "text": "再点击顶部的知识库",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 710.075,
      "end": 711.935,
      "text": "最后点击右上角的创建",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 711.935,
      "end": 714.155,
      "text": "名称填写家庭AI测试库",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 714.155,
      "end": 715.415,
      "text": "目标可以简单写真",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 715.415,
      "end": 718.655,
      "text": "用于测试本地RIG文档检索和应用功能",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 718.655,
      "end": 720.035,
      "text": "知识库创建完之后",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 720.035,
      "end": 721.415,
      "text": "再点击右上角的加号",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 721.415,
      "end": 722.415,
      "text": "选择上传文件",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 722.415,
      "end": 724.935,
      "text": "把刚才的txt文件上传进去",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 724.935,
      "end": 727.735,
      "text": "后面也可以上传带有文字层的PDF",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 727.735,
      "end": 729.695,
      "text": "Docx或Markdown文档",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 729.695,
      "end": 730.615,
      "text": "上乘完之后",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 730.615,
      "end": 732.335,
      "text": "在新兑化的聊天输入框里",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 732.335,
      "end": 734.695,
      "text": "通过警号手动附加这个支持库",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 734.695,
      "end": 736.735,
      "text": "而我这里为了教程演示方便",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 736.735,
      "end": 737.655,
      "text": "先手动选择",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 737.655,
      "end": 739.695,
      "text": "因为小参数模型主动判断",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 739.695,
      "end": 740.735,
      "text": "什么时候动用支持库",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 740.735,
      "end": 741.875,
      "text": "有时候不够稳定",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 741.875,
      "end": 743.675,
      "text": "然后我们可以问一个问题",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 743.675,
      "end": 745.135,
      "text": "只依据支持库回答",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 745.135,
      "end": 746.935,
      "text": "项目代号和设备维护时间",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 746.935,
      "end": 747.655,
      "text": "氛围是什么",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 747.655,
      "end": 748.815,
      "text": "请给出引用",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 748.815,
      "end": 750.035,
      "text": "正常情况下",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 750.035,
      "end": 750.975,
      "text": "耐心等待一会儿",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 750.975,
      "end": 753.475,
      "text": "他应该回答项目代号是清零",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 753.475,
      "end": 755.335,
      "text": "设备维护时间是每月第一",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 755.335,
      "end": 756.355,
      "text": "周六上午十点",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 756.355,
      "end": 759.055,
      "text": "而且回答旁边能够看到对应的资料引用",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 759.055,
      "end": 762.895,
      "text": "那么这套最基础的REG知识库就算大奖成功了",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 762.895,
      "end": 763.975,
      "text": "不过这样配置下来",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 763.975,
      "end": 766.095,
      "text": "你可能会发现还有一个地方容易踩坑",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 766.095,
      "end": 767.875,
      "text": "如果你的PDF是扫描件",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 767.875,
      "end": 770.055,
      "text": "本身没有可以复制的文字层",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 770.055,
      "end": 772.015,
      "text": "就不要拿它来测试REG了",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 772.015,
      "end": 774.775,
      "text": "因为知识库首先需要能够读取里面的文字",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 774.775,
      "end": 776.035,
      "text": "才能进行后续的处理",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 776.035,
      "end": 778.315,
      "text": "所以这种文件就需要先经过OCR",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 778.315,
      "end": 781.335,
      "text": "把图片里的文字设备成TXC或者Markdown",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 781.335,
      "end": 783.275,
      "text": "再上传到知识库来进行处理",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 783.275,
      "end": 786.255,
      "text": "接下来我们就继续把本地OCR的功能",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 786.255,
      "end": 787.235,
      "text": "也接近来",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 2,
      "language_run": 9,
      "source": "whisper"
    },
    {
      "start": 792.446,
      "end": 794.796,
      "text": "前面我们已经让AI能够读取次的文档",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 794.796,
      "end": 796.976,
      "text": "接下来再补上图片理解和OCR",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 796.976,
      "end": 799.816,
      "text": "也就是让这台本地AI不只能读文字",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 799.816,
      "end": 800.956,
      "text": "还可以看图片",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 800.956,
      "end": 802.856,
      "text": "那这里先区分两个概念",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 802.856,
      "end": 805.436,
      "text": "普通图片理解更适合看一张图里有什么",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 805.436,
      "end": 806.696,
      "text": "界面再显示什么",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 806.696,
      "end": 808.496,
      "text": "再顺便识别其中的文字",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 808.496,
      "end": 811.496,
      "text": "OCR则更专注于把图片里的文字",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 811.496,
      "end": 814.116,
      "text": "表格或者图表内容尽可能提取出来",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 814.116,
      "end": 816.976,
      "text": "先在OpenWeb UI里选择我们前面使用的",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 816.976,
      "end": 819.316,
      "text": "千万305 4B-KL4KM",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 819.316,
      "end": 820.816,
      "text": "如果你的Mac配置更高",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 820.816,
      "end": 822.016,
      "text": "也可以用9B的版本",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 822.016,
      "end": 824.916,
      "text": "上传一张同时包含物体和文字图片",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 824.916,
      "end": 825.596,
      "text": "然后输入",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 825.596,
      "end": 826.796,
      "text": "先描述画面内容",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 826.796,
      "end": 828.516,
      "text": "再逐项提取可见文字",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 828.516,
      "end": 830.496,
      "text": "不确定的部分标注为不确定",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 830.496,
      "end": 831.596,
      "text": "不需要自信补圈",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 831.596,
      "end": 833.736,
      "text": "这里可以分别拿产品包装",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 833.736,
      "end": 834.716,
      "text": "软件界面截图",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 834.716,
      "end": 835.456,
      "text": "中文表格",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 835.456,
      "end": 837.436,
      "text": "或者已经做好隐私透明的发票",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 837.436,
      "end": 839.136,
      "text": "和数据照片来测试一下",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 839.136,
      "end": 841.116,
      "text": "这样可以比较直观的判断出",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 841.116,
      "end": 842.956,
      "text": "模型到底是理解了图片内容",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 842.956,
      "end": 845.396,
      "text": "还是只是把里面的文字给识别出来",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 845.396,
      "end": 847.476,
      "text": "如果你出现了类似我这样的提示",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 847.476,
      "end": 849.156,
      "text": "就前往OpenWeb UI的设置",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 849.156,
      "end": 849.716,
      "text": "模型",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 849.716,
      "end": 851.556,
      "text": "找到刚才你前面使用的模型",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 852.016,
      "end": 852.656,
      "text": "编辑界面",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 852.656,
      "end": 855.296,
      "text": "找到高级参数下的NUMCTX",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 855.296,
      "end": 855.696,
      "text": "Olama",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 855.696,
      "end": 857.456,
      "text": "针对我16GB的Mac mini",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 857.456,
      "end": 859.096,
      "text": "将数字更改为8192",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 859.096,
      "end": 861.156,
      "text": "然后重复上的操作就可以了",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 861.156,
      "end": 862.896,
      "text": "如果我们的目标不是理解图片",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 862.896,
      "end": 865.376,
      "text": "而是尽可能把图片里的文字给提取出来",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 865.376,
      "end": 868.536,
      "text": "就可以使用我们前面已经下载好的GLM OCR",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 868.536,
      "end": 870.036,
      "text": "OCR模型运行前",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 870.036,
      "end": 872.276,
      "text": "如果较大的千万模型还占着同一印存",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 872.276,
      "end": 873.636,
      "text": "先可以把它停掉",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 873.636,
      "end": 875.856,
      "text": "接下来我们准备好需要识别图片",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 875.856,
      "end": 877.456,
      "text": "放到一个统一的文件夹中",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 877.456,
      "end": 878.976,
      "text": "比如我这个目录像",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 878.976,
      "end": 881.256,
      "text": "对于我们Mac仿达中的用户文件夹",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 881.256,
      "end": 882.216,
      "text": "AI test",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 882.216,
      "end": 884.636,
      "text": "这里手动新建一个test的文件夹就可以了",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 884.636,
      "end": 886.616,
      "text": "对于需要识别不同的内容图片",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 886.616,
      "end": 888.676,
      "text": "比如普通收据表格图表",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 888.676,
      "end": 890.096,
      "text": "都可以使用不同的命令",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 890.096,
      "end": 891.756,
      "text": "注意命令结尾的文件名",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 891.756,
      "end": 893.776,
      "text": "需要换成你自己图片的文件名",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 893.776,
      "end": 894.876,
      "text": "第一次测试时",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 894.876,
      "end": 896.576,
      "text": "建议每次只处理一张图片",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 896.576,
      "end": 898.916,
      "text": "先确认整套工作流能够稳定运行",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 898.916,
      "end": 900.956,
      "text": "如果遇到扫描PDF",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 900.956,
      "end": 903.876,
      "text": "也不用一开始就折腾完整的PDF解析工具",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 903.876,
      "end": 905.276,
      "text": "最简单办法就是",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 905.276,
      "end": 907.476,
      "text": "先用Microx自带的预览打开PDF",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 907.476,
      "end": 909.176,
      "text": "把其中一页导成PN器",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 909.176,
      "end": 911.496,
      "text": "再交给GLM OCR进行处理",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 911.496,
      "end": 913.676,
      "text": "等确认单页OCR没问题后",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 913.676,
      "end": 915.796,
      "text": "再考虑批量处理整份文档",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 915.796,
      "end": 917.996,
      "text": "识别出来的内容还可以继续利用",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 917.996,
      "end": 919.976,
      "text": "保存成TXT或者Markdown",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 919.976,
      "end": 922.576,
      "text": "再上传到我们上一节创建的知识库",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 922.576,
      "end": 924.216,
      "text": "然后交给签问做摘要",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 924.216,
      "end": 926.136,
      "text": "自端整理或者检索问答",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 926.136,
      "end": 928.056,
      "text": "那么整个流程就传起来了",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 928.056,
      "end": 930.996,
      "text": "如果OCR出现空白重复或者乱码",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 930.996,
      "end": 932.196,
      "text": "可以先更新Olam",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 932.196,
      "end": 934.276,
      "text": "然后每次只处理一张图片",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 934.276,
      "end": 935.436,
      "text": "图片过大时",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 935.436,
      "end": 939.016,
      "text": "可以把长辨锁到大约1600至2200像素",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 939.016,
      "end": 941.156,
      "text": "并优先使用PNG或者GPG",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 941.156,
      "end": 943.976,
      "text": "同时检查原图有没有明显模糊",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 943.976,
      "end": 945.056,
      "text": "倾斜反光",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 945.056,
      "end": 947.056,
      "text": "这些都会直接影响识别效果",
      "language": "zh",
      "chunk": 1,
      "local_language_run": 3,
      "language_run": 10,
      "source": "whisper"
    },
    {
      "start": 952.24,
      "end": 954.07,
      "text": "除了文字和图片之外",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 954.07,
      "end": 955.95,
      "text": "本地还有一个非常实用的场景",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 955.95,
      "end": 957.55,
      "text": "就是处理视频和录音",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 957.55,
      "end": 959.41,
      "text": "比如把会议录音转成文字",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 959.41,
      "end": 960.91,
      "text": "把视频自动生成字幕",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 960.91,
      "end": 964.21,
      "text": "或者把自己的素材库快速整理成可搜索的文本",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 964.21,
      "end": 966.19,
      "text": "这里我们使用Whisper.cpp",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 966.19,
      "end": 968.59,
      "text": "它是Whisper模型C++实现的",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 968.59,
      "end": 970.61,
      "text": "针对Apple Silicon提供Mental加速",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 970.61,
      "end": 973.35,
      "text": "可以直接利用Mac的GPU进行本地转写",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 973.35,
      "end": 975.23,
      "text": "首先安装需要用到的工具",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 975.23,
      "end": 977.83,
      "text": "FMMP1G负责处理音频格式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 977.83,
      "end": 980.15,
      "text": "Whisper.cpp负责语音识别",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 980.15,
      "end": 982.09,
      "text": "如果你的Mac已经安装Homebrew",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 982.09,
      "end": 983.93,
      "text": "可以直接输入代码来进行安装",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 983.93,
      "end": 985.67,
      "text": "如果你仍然没有安装Homebrew",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 985.67,
      "end": 988.31,
      "text": "安装whisper.cpp可以走圆码编译路线",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 988.31,
      "end": 989.79,
      "text": "不过对于普通用户来说",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 989.79,
      "end": 991.81,
      "text": "这一步会比Homebrew更复杂一些",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 991.81,
      "end": 994.35,
      "text": "还需要提前安装Xcode的Command Align Tools",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 994.35,
      "end": 995.69,
      "text": "这里就不展开了",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 995.69,
      "end": 996.63,
      "text": "安装完成后",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 996.63,
      "end": 999.13,
      "text": "可以通过下面两个命令来确认是否正常",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 999.13,
      "end": 1000.73,
      "text": "接下来下载Whisper模型",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1000.73,
      "end": 1002.19,
      "text": "我这里选择的是这个",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1002.19,
      "end": 1004.57,
      "text": "这是一个支持中文和多语言的模型",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1004.57,
      "end": 1006.83,
      "text": "相比完整Large模型体积更小",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1006.83,
      "end": 1008.71,
      "text": "同时保留了比较好的转眼效果",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1008.71,
      "end": 1010.15,
      "text": "更适合Mac本地运行",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1010.15,
      "end": 1013.17,
      "text": "我们先创建存放模型和临时文件的目录",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1013.17,
      "end": 1014.19,
      "text": "然后下载模型",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1014.19,
      "end": 1017.23,
      "text": "下载完整后再计算一次SHA-1的值",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1017.23,
      "end": 1021.15,
      "text": "方便和whisper.cpp模型文档公开的接应值进行核对",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1021.15,
      "end": 1023.55,
      "text": "模型准备好后就可以开始处理视频",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1023.55,
      "end": 1026.27,
      "text": "假设你的视频文件位于这个文件夹",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1026.27,
      "end": 1029.33,
      "text": "也就是MicroOS用户目录的影片文件夹当中",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1029.33,
      "end": 1032.07,
      "text": "大家根据自己的实际文件存储位置进行修改",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1032.07,
      "end": 1037.05,
      "text": "首先使用FMMP1G把视频里的音频转换成Whisper更适合处理的格式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1037.05,
      "end": 1038.89,
      "text": "这几个参数简单介绍一下",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1038.89,
      "end": 1040.43,
      "text": "VM表示去掉视频",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1040.43,
      "end": 1041.25,
      "text": "只保留音频",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1041.25,
      "end": 1044.63,
      "text": "AR16表示转换成16000赫兹的采样率",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1044.63,
      "end": 1046.31,
      "text": "AC1表示单身道",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1046.31,
      "end": 1049.23,
      "text": "PCM S16LE则是转换成Whisper",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1049.23,
      "end": 1051.07,
      "text": "更容易处理的WAB格式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1051.07,
      "end": 1052.33,
      "text": "接下来运行Whisper",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1052.33,
      "end": 1054.37,
      "text": "同时生成文字稿和字幕文件",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1054.37,
      "end": 1057.27,
      "text": "-M用来指定我们前面下载的Whisper模型",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1057.27,
      "end": 1059.89,
      "text": "-F指定需要转写的音频文件",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1059.89,
      "end": 1061.67,
      "text": "-LZH告诉我们模型",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1061.67,
      "end": 1063.49,
      "text": "这段文字主要内容是中文",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1063.49,
      "end": 1065.73,
      "text": "-Prompt则可以提前提供一些",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1065.73,
      "end": 1067.37,
      "text": "容易识别错误的中有名词",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1067.37,
      "end": 1068.31,
      "text": "比如Mac mini",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1068.31,
      "end": 1069.19,
      "text": "Apple Silicon",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1069.19,
      "end": 1070.41,
      "text": "Olama和千问",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1070.41,
      "end": 1071.93,
      "text": "对于科技类视频有其有用",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1071.93,
      "end": 1074.19,
      "text": "比如这里我就加上了我的Lookup Weekly",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1074.19,
      "end": 1078.75,
      "text": "后面的-otxt-oosrt和-ovtt",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1078.75,
      "end": 1080.95,
      "text": "分别让它同时生成TXT文字稿",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1080.95,
      "end": 1083.31,
      "text": "SRT字幕和VTT字幕",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1083.31,
      "end": 1084.69,
      "text": "最后通过-off",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1084.69,
      "end": 1086.77,
      "text": "指定输出文件的位置和名称",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1086.77,
      "end": 1089.03,
      "text": "运行完成后会生成这些文件",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1089.03,
      "end": 1090.89,
      "text": "其中TXT是纯文字稿",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1090.89,
      "end": 1093.43,
      "text": "适合后续交给AI总结或者整理",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1093.43,
      "end": 1095.33,
      "text": "SRT是常见的字幕格式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1095.33,
      "end": 1096.81,
      "text": "可以直接导入紧接软件",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1096.81,
      "end": 1099.75,
      "text": "VTT是主要用于网页播放器的字幕",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1099.75,
      "end": 1101.05,
      "text": "如果是8GB的Mac",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1101.05,
      "end": 1105.73,
      "text": "也可以先尝试这里的Large-V3-Turbo-Q50电话模型",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1105.73,
      "end": 1108.67,
      "text": "如果实际转写时内存压力较大或者速度不理想",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1108.67,
      "end": 1110.97,
      "text": "再换成体积更小的Small多语言模型",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1110.97,
      "end": 1113.43,
      "text": "不要下载来带EAN后置的英文专用模型",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1113.43,
      "end": 1115.09,
      "text": "因为它只针对英文优化",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1115.09,
      "end": 1116.41,
      "text": "如果需要中文转写",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1116.41,
      "end": 1117.71,
      "text": "也不要开启翻译参数",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1117.71,
      "end": 1119.71,
      "text": "否则模型会把内容翻译成英文",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1119.71,
      "end": 1121.31,
      "text": "而不是保留原始的语言",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1121.31,
      "end": 1124.43,
      "text": "到这里整套本地i工作流基本已经闭环",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1124.43,
      "end": 1126.01,
      "text": "视频和录音可以通过Whisper",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1126.01,
      "end": 1126.77,
      "text": "转写文字",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1126.77,
      "end": 1128.21,
      "text": "再交给前文做总结",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1128.21,
      "end": 1128.65,
      "text": "整理",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1128.65,
      "end": 1130.87,
      "text": "最后也可以存入前面搭建的知识库",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1130.87,
      "end": 1133.17,
      "text": "实现自己本地的内容管理系统",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 1,
      "language_run": 11,
      "source": "whisper"
    },
    {
      "start": 1138.354,
      "end": 1143.264,
      "text": "前面我们已经分别完成了本地云转写OpenWebUIAPI和签问调用",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1143.264,
      "end": 1145.844,
      "text": "至于使用N8N搭建一条自动化工作流",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1145.844,
      "end": 1147.504,
      "text": "上传音频或者视频后",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1147.504,
      "end": 1149.424,
      "text": "N8N自动调用Whisper来转写",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1149.424,
      "end": 1151.764,
      "text": "再把转写稿交给签问来整理",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1151.764,
      "end": 1155.084,
      "text": "最后返回摘要提纲标题和证捷建议",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1155.084,
      "end": 1156.844,
      "text": "不需要每次手动复制粘贴",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1156.844,
      "end": 1159.124,
      "text": "首先把前面安装好的Whisper.cpp",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1159.124,
      "end": 1161.744,
      "text": "变成一个可以被其他程序调用的本地API",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1161.744,
      "end": 1164.224,
      "text": "先尝试允许Mac本机访问",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1164.224,
      "end": 1165.804,
      "text": "保持这个中端窗口开启",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1165.804,
      "end": 1168.164,
      "text": "再新开一个中南app窗口进行测试",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1168.164,
      "end": 1170.204,
      "text": "如果能够正常返回转写结果",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1170.204,
      "end": 1172.064,
      "text": "Whisper API就已经运行成功",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1172.064,
      "end": 1174.684,
      "text": "后面N8N会运行在Docker容器里",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1174.684,
      "end": 1176.684,
      "text": "如果出现无法连接Whisper的情况",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1176.684,
      "end": 1178.484,
      "text": "可以按Ctrl加C停止服务",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1178.484,
      "end": 1179.384,
      "text": "再这样改",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1179.384,
      "end": 1180.504,
      "text": "这里需要特别注意",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1180.504,
      "end": 1183.944,
      "text": "Whisper server本身没有用户登录和全线验证的功能",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1183.944,
      "end": 1186.044,
      "text": "所以使用0.0.0.0时",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1186.044,
      "end": 1188.364,
      "text": "只建议放在可信的家庭局网内",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1188.364,
      "end": 1189.964,
      "text": "不要进行路由器端口转发",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1189.964,
      "end": 1192.704,
      "text": "也不要把8178端口直接暴露到公网",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1192.704,
      "end": 1193.904,
      "text": "接下来安装N8N",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1193.904,
      "end": 1196.004,
      "text": "这里使用Docker Desktop来运行",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1196.004,
      "end": 1197.444,
      "text": "并固定N8N版本",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1197.444,
      "end": 1201.144,
      "text": "避免后续更新导致教程界面或者节点配置发生变化",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1201.144,
      "end": 1203.844,
      "text": "先从Docker官网下载安装Docker Desktop",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1203.844,
      "end": 1205.984,
      "text": "偷动到应用程序无电夹并启动",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1205.984,
      "end": 1207.364,
      "text": "确认Docker正常运行后",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1207.364,
      "end": 1209.544,
      "text": "再打开终端App输入下面的命令",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 2,
      "language_run": 12,
      "source": "whisper"
    },
    {
      "start": 1216.852,
      "end": 1219.162,
      "text": "接下来生成一段随机的加密密钥",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 3,
      "language_run": 13,
      "source": "whisper"
    },
    {
      "start": 1222.856,
      "end": 1225.186,
      "text": "把生成的结果替换到下面的XXX",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 4,
      "language_run": 14,
      "source": "whisper"
    },
    {
      "start": 1229.721,
      "end": 1231.851,
      "text": "然后我们访问127.0.0.1",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 5,
      "language_run": 15,
      "source": "whisper"
    },
    {
      "start": 1231.851,
      "end": 1232.891,
      "text": "号5678",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 5,
      "language_run": 15,
      "source": "whisper"
    },
    {
      "start": 1232.891,
      "end": 1236.111,
      "text": "第一次进入需要创建N8N的本地管理员账号",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 5,
      "language_run": 15,
      "source": "whisper"
    },
    {
      "start": 1236.111,
      "end": 1238.551,
      "text": "接下来我们来新建一条workflow工作流",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 5,
      "language_run": 15,
      "source": "whisper"
    },
    {
      "start": 1238.551,
      "end": 1240.491,
      "text": "这套流程一共只有四个节点",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 5,
      "language_run": 15,
      "source": "whisper"
    },
    {
      "start": 1240.491,
      "end": 1241.046,
      "text": "Webhook",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 5,
      "language_run": 15,
      "source": "whisper"
    },
    {
      "start": 1241.046,
      "end": 1243.816,
      "text": "webhook HTTP request with perjuanxie",
      "language": "en",
      "chunk": 2,
      "local_language_run": 6,
      "language_run": 16,
      "source": "whisper"
    },
    {
      "start": 1243.816,
      "end": 1244.346,
      "text": "HTTP request",
      "language": "en",
      "chunk": 2,
      "local_language_run": 6,
      "language_run": 16,
      "source": "whisper"
    },
    {
      "start": 1244.346,
      "end": 1245.656,
      "text": "HTTP Request的签文整理",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 7,
      "language_run": 17,
      "source": "whisper"
    },
    {
      "start": 1245.656,
      "end": 1247.216,
      "text": "Respond to Webhook",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 7,
      "language_run": 17,
      "source": "whisper"
    },
    {
      "start": 1247.216,
      "end": 1248.896,
      "text": "第一个是Webhook节点",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 7,
      "language_run": 17,
      "source": "whisper"
    },
    {
      "start": 1248.896,
      "end": 1250.596,
      "text": "它相当于整条工作流的入口",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 7,
      "language_run": 17,
      "source": "whisper"
    },
    {
      "start": 1250.596,
      "end": 1252.896,
      "text": "以后只要把视频post到这个地址",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 7,
      "language_run": 17,
      "source": "whisper"
    },
    {
      "start": 1252.896,
      "end": 1254.356,
      "text": "工作流就会自动开始",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 7,
      "language_run": 17,
      "source": "whisper"
    },
    {
      "start": 1254.356,
      "end": 1256.056,
      "text": "Webhook的具体设置如下",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 7,
      "language_run": 17,
      "source": "whisper"
    },
    {
      "start": 1256.056,
      "end": 1258.036,
      "text": "没提到的设置保持不便即可",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 7,
      "language_run": 17,
      "source": "whisper"
    },
    {
      "start": 1268.267,
      "end": 1270.297,
      "text": "第二个节点添加http request",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1270.297,
      "end": 1272.417,
      "text": "名称可以改成whisper转写",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1272.417,
      "end": 1276.417,
      "text": "它负责把刚刚上传的音视频文件交给Mac上的whisper server",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1276.417,
      "end": 1279.117,
      "text": "第二个节点http request的具体设置如下",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1279.117,
      "end": 1280.997,
      "text": "没提到的设置保持不便即可",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1280.997,
      "end": 1284.197,
      "text": "值得注意的是n8nbinaryfile这类型中",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1284.197,
      "end": 1289.417,
      "text": "input data field name需要填写webhook节点实际收到的二进制字段名称",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1289.417,
      "end": 1291.097,
      "text": "点击前一个webhook节点",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1291.097,
      "end": 1293.557,
      "text": "在output的binary中进行查看",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1293.557,
      "end": 1295.797,
      "text": "如果显示为data那就填写为data",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1295.797,
      "end": 1297.077,
      "text": "如果显示为Data0",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1297.077,
      "end": 1298.357,
      "text": "那就填写为Data0",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1298.357,
      "end": 1299.857,
      "text": "这个名称必须完全一致",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1299.857,
      "end": 1300.857,
      "text": "否则会出现",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1300.857,
      "end": 1303.337,
      "text": "The item has no binary field的报错",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1303.337,
      "end": 1305.877,
      "text": "第三个节点继续添加http request",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1305.877,
      "end": 1307.497,
      "text": "名称改成前文整理",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1307.497,
      "end": 1310.617,
      "text": "这个节点负责把Whisper返回完整转写稿",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1310.617,
      "end": 1313.097,
      "text": "交给前面搭建好的OpenWeb UI API",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1313.097,
      "end": 1315.937,
      "text": "第三个节点http request具体设置如下",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1315.937,
      "end": 1317.957,
      "text": "没提到的设置保持不变即可",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 8,
      "language_run": 18,
      "source": "whisper"
    },
    {
      "start": 1328.105,
      "end": 1330.095,
      "text": "接下来回到前面的JSON输入框",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 9,
      "language_run": 19,
      "source": "whisper"
    },
    {
      "start": 1330.095,
      "end": 1332.295,
      "text": "把整个字段切换到expression模式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 9,
      "language_run": 19,
      "source": "whisper"
    },
    {
      "start": 1332.295,
      "end": 1333.875,
      "text": "再粘贴下面的表达式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 9,
      "language_run": 19,
      "source": "whisper"
    },
    {
      "start": 1333.875,
      "end": 1335.155,
      "text": "它的逻辑很简单",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 9,
      "language_run": 19,
      "source": "whisper"
    },
    {
      "start": 1335.155,
      "end": 1338.175,
      "text": "N8N自动读取Whisper转写节点里面的text",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 9,
      "language_run": 19,
      "source": "whisper"
    },
    {
      "start": 1338.175,
      "end": 1340.775,
      "text": "再连同预设提示词一起发给亲吻",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 9,
      "language_run": 19,
      "source": "whisper"
    },
    {
      "start": 1343.996,
      "end": 1346.026,
      "text": "最后添加Respound to Webhook节点",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 10,
      "language_run": 20,
      "source": "whisper"
    },
    {
      "start": 1346.026,
      "end": 1349.006,
      "text": "它负责把整条工作流结果返回给定用者",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 10,
      "language_run": 20,
      "source": "whisper"
    },
    {
      "start": 1349.006,
      "end": 1350.746,
      "text": "Respound with选择Json",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 10,
      "language_run": 20,
      "source": "whisper"
    },
    {
      "start": 1350.746,
      "end": 1353.346,
      "text": "然后把Respound body切换成Expression模式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 10,
      "language_run": 20,
      "source": "whisper"
    },
    {
      "start": 1353.346,
      "end": 1354.566,
      "text": "粘贴下面的表达式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 10,
      "language_run": 20,
      "source": "whisper"
    },
    {
      "start": 1354.566,
      "end": 1356.526,
      "text": "Respound code保持200即可",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 10,
      "language_run": 20,
      "source": "whisper"
    },
    {
      "start": 1366.204,
      "end": 1367.494,
      "text": "工作流设置完成以后",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 11,
      "language_run": 21,
      "source": "whisper"
    },
    {
      "start": 1367.494,
      "end": 1368.894,
      "text": "先不要急着正式启用",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 11,
      "language_run": 21,
      "source": "whisper"
    },
    {
      "start": 1368.894,
      "end": 1371.734,
      "text": "在N8N里点击Listen for Test Event",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 11,
      "language_run": 21,
      "source": "whisper"
    },
    {
      "start": 1371.734,
      "end": 1374.094,
      "text": "让Webhook等待一次测试请求",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 11,
      "language_run": 21,
      "source": "whisper"
    },
    {
      "start": 1374.094,
      "end": 1376.134,
      "text": "然后回到终端执行下面的命令",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 11,
      "language_run": 21,
      "source": "whisper"
    },
    {
      "start": 1376.134,
      "end": 1378.814,
      "text": "注意要把XAI Token后面的章位内容",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 11,
      "language_run": 21,
      "source": "whisper"
    },
    {
      "start": 1378.814,
      "end": 1381.454,
      "text": "替换成你前面实际设置的随机自如串",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 11,
      "language_run": 21,
      "source": "whisper"
    },
    {
      "start": 1387.735,
      "end": 1388.825,
      "text": "如果整个流程正常",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 12,
      "language_run": 22,
      "source": "whisper"
    },
    {
      "start": 1388.825,
      "end": 1390.205,
      "text": "你最终应该会得到",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 12,
      "language_run": 22,
      "source": "whisper"
    },
    {
      "start": 1390.205,
      "end": 1391.145,
      "text": "类似这样的结果",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 12,
      "language_run": 22,
      "source": "whisper"
    },
    {
      "start": 1394.264,
      "end": 1397.334,
      "text": "测试成功以后回到N8N的工作流程页面",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1397.334,
      "end": 1399.294,
      "text": "点击右上角的Publish发布",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1399.294,
      "end": 1403.094,
      "text": "发布后就不需要每次先点击Lessens for Test Event了",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1403.094,
      "end": 1406.594,
      "text": "正式Webhook地址也会从Webhook Test AI Video Notes",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1406.594,
      "end": 1408.754,
      "text": "变成Webhook AI Video Notes",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1408.754,
      "end": 1410.754,
      "text": "以后只要把视频传到这个接口",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1410.754,
      "end": 1413.074,
      "text": "整套流程就会自动变成Whisper",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1413.074,
      "end": 1415.074,
      "text": "在本地把视频转成文字",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1415.074,
      "end": 1417.194,
      "text": "千万再在本地完成摘要",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1417.194,
      "end": 1418.974,
      "text": "提纲 标题和章节建议",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1418.974,
      "end": 1421.694,
      "text": "最后把转写稿和整理结果一起返回",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1421.694,
      "end": 1423.654,
      "text": "整个过程中的语音和文本内容",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1423.654,
      "end": 1425.694,
      "text": "都可以在自己的Mac上完成处理",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1425.694,
      "end": 1428.094,
      "text": "到这里整套工作流已经跑通了",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1428.094,
      "end": 1430.594,
      "text": "不过直接使用Curve调用正式外部后格式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1430.594,
      "end": 1432.554,
      "text": "中端反围的是标准的JSON",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1432.554,
      "end": 1434.514,
      "text": "所以会看到大量的斜杠N",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1434.514,
      "end": 1436.114,
      "text": "引号和转译字符",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1436.114,
      "end": 1437.794,
      "text": "比如斜杠N其实就是换",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1437.794,
      "end": 1438.774,
      "text": "并不是烂码",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1438.774,
      "end": 1442.514,
      "text": "只是JSON保存Markdown和多项文字时的正常格式",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1442.514,
      "end": 1444.654,
      "text": "为让最终结果更加适合阅读",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1444.654,
      "end": 1447.214,
      "text": "我们可以再用JQ把JSON里的转写稿",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1447.214,
      "end": 1448.914,
      "text": "和签文之力结果提取出来",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1448.914,
      "end": 1450.274,
      "text": "如果已经安装Homebrew",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1450.274,
      "end": 1452.314,
      "text": "则可以执行Brew install GQ",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1452.314,
      "end": 1453.414,
      "text": "安装完整后",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1453.414,
      "end": 1456.094,
      "text": "把刚才电用正式工作流的命令稍微改一下",
      "language": "zh",
      "chunk": 2,
      "local_language_run": 13,
      "language_run": 23,
      "source": "whisper"
    },
    {
      "start": 1462.315,
      "end": 1464.305,
      "text": "如果只是想看签问整理后的结果",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 1,
      "language_run": 24,
      "source": "whisper"
    },
    {
      "start": 1464.305,
      "end": 1465.445,
      "text": "也可以进一步讲话",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 1,
      "language_run": 24,
      "source": "whisper"
    },
    {
      "start": 1468.32,
      "end": 1470.37,
      "text": "如果只需要Whisper的完整转眼稿",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 2,
      "language_run": 25,
      "source": "whisper"
    },
    {
      "start": 1470.37,
      "end": 1471.35,
      "text": "也可以这么改",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 2,
      "language_run": 25,
      "source": "whisper"
    },
    {
      "start": 1474.326,
      "end": 1476.336,
      "text": "哦对了如果你的Mac没有安装Homebrew",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 3,
      "language_run": 26,
      "source": "whisper"
    },
    {
      "start": 1476.336,
      "end": 1478.376,
      "text": "也不一定非要额外安装JQ",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 3,
      "language_run": 26,
      "source": "whisper"
    },
    {
      "start": 1478.376,
      "end": 1480.396,
      "text": "前面已经准备了Python 3.1",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 3,
      "language_run": 26,
      "source": "whisper"
    },
    {
      "start": 1480.396,
      "end": 1483.876,
      "text": "可以直接用Python自带的JSON模块完成同样的处理",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 3,
      "language_run": 26,
      "source": "whisper"
    },
    {
      "start": 1489.353,
      "end": 1491.983,
      "text": "视频最后在补一个实际使用时一定会碰到的问题",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1491.983,
      "end": 1493.763,
      "text": "Mac mini重启以后",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1493.763,
      "end": 1495.443,
      "text": "哪些东西需要重新打开",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1495.443,
      "end": 1498.323,
      "text": "我们前面搭建的模型和数据都保存在硬盘里",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1498.323,
      "end": 1500.183,
      "text": "不需要重新安装或者重新下载",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1500.183,
      "end": 1505.523,
      "text": "但Olama,OpenWeb UI,Whisper Server和N8N这些运行中的服务需要重新启动",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1505.523,
      "end": 1507.263,
      "text": "建议按照下面这个顺序来",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1507.263,
      "end": 1509.143,
      "text": "首先启动最底层的Olama",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1509.143,
      "end": 1511.003,
      "text": "如果安装的是MicOS版的Olama",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1511.003,
      "end": 1512.743,
      "text": "直接打开Olama app即可",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1512.743,
      "end": 1515.583,
      "text": "也可以在中端里执行Open-AOlama",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1515.583,
      "end": 1518.143,
      "text": "然后可以检查Olama是否已经恢复",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1518.143,
      "end": 1520.903,
      "text": "如果能够正常返回Olama的版本信息",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1520.903,
      "end": 1523.143,
      "text": "就说明底层模型服务已经正常",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1523.143,
      "end": 1525.083,
      "text": "接下来再来启动OpenWeb UI",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1525.083,
      "end": 1528.903,
      "text": "充计后可以先确认MacMe的局域网IP地址有没有发生变化",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1528.903,
      "end": 1531.863,
      "text": "因为我们前面的启动命令里填写过这个地址",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1531.863,
      "end": 1533.543,
      "text": "如果地址没有发生变化",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1533.543,
      "end": 1534.763,
      "text": "就继续使用之前的IP",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1534.763,
      "end": 1535.863,
      "text": "如果变化了",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1535.863,
      "end": 1537.783,
      "text": "下面命令行里的也要一起替换",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1537.783,
      "end": 1539.643,
      "text": "这个中端装口需要保持运行",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1539.643,
      "end": 1540.883,
      "text": "然后617访问",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1540.883,
      "end": 1542.843,
      "text": "127.0.0.1网号3000",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1542.843,
      "end": 1544.203,
      "text": "局域网中的其他设备",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1544.203,
      "end": 1546.463,
      "text": "仍然访问MacMe的IP地址网号3000",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1546.463,
      "end": 1550.643,
      "text": "长期使用的话也可以在路由器里给Mac mini固定一个局网地址",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1550.643,
      "end": 1552.763,
      "text": "选择每次重建后来进行检查",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1552.763,
      "end": 1555.483,
      "text": "如果只是聊天知识库或者OpenWeb UI",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1555.483,
      "end": 1557.443,
      "text": "其实不用一直启动Whisper Server",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1557.443,
      "end": 1561.983,
      "text": "只有需要运行视频转写或者后面的N8n视频工作流时再打开即可",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1561.983,
      "end": 1563.983,
      "text": "同样这个中端创口需要保持运行",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1563.983,
      "end": 1565.703,
      "text": "可以打开另一个中端测试",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1565.703,
      "end": 1567.483,
      "text": "只要服务器能够这样响应",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1567.483,
      "end": 1569.703,
      "text": "就可以继续交给N8n来调用了",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1569.703,
      "end": 1571.763,
      "text": "接着打开Docker Desktop App",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1571.763,
      "end": 1576.083,
      "text": "因为我们创建N8n容器时已经设置了Restart Unless Stoped",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1576.083,
      "end": 1577.143,
      "text": "所以正常情况下",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1577.443,
      "end": 1578.543,
      "text": "只要Docker app启动",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1578.543,
      "end": 1580.723,
      "text": "之前的N8n容器也会自动恢复运行",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1580.883,
      "end": 1582.883,
      "text": "不需要重新执行那一串Docker run",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1583.283,
      "end": 1584.883,
      "text": "可以先检查Docker PS",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1584.883,
      "end": 1586.523,
      "text": "如果Liviar里能够看到N8n",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1586.523,
      "end": 1588.483,
      "text": "就说明N8n已经自动恢复",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1588.683,
      "end": 1589.483,
      "text": "如果没有启动",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1589.483,
      "end": 1591.783,
      "text": "可以手动执行Docker StartN8n",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1591.943,
      "end": 1595.383,
      "text": "然后留言器打开127.0.0.1-5678",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1595.823,
      "end": 1596.983,
      "text": "如果能够进入N8n",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1596.983,
      "end": 1599.143,
      "text": "就说明自动化服务已经恢复正常",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1599.243,
      "end": 1602.283,
      "text": "到这里我们就从一台普通的Mac mini开始",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1602.283,
      "end": 1605.443,
      "text": "完成一套相对完整的家庭本地AI服务",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1605.443,
      "end": 1607.443,
      "text": "Olama负责运行AI模型",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1607.443,
      "end": 1610.443,
      "text": "OpenWeb UI提供对话界面和统一API",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1610.443,
      "end": 1613.303,
      "text": "Embedding和RAG负责自己的文档知识库",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1613.303,
      "end": 1616.223,
      "text": "GLM OCR处理图片和扫描件",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1616.223,
      "end": 1618.823,
      "text": "Whisper.cpp负责视频录音转写",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1618.823,
      "end": 1623.003,
      "text": "再通过N8N把这些能力串成一套简单的自动化工作流",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1623.003,
      "end": 1626.503,
      "text": "这套方案其实不只是在Mac上跑了一个AI模型",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1626.503,
      "end": 1631.003,
      "text": "当对话、文档、图片、映视频和自动化都连起来之后",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1631.003,
      "end": 1633.883,
      "text": "这台Mac mini才开始从一台跑模型的电脑",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1633.883,
      "end": 1636.123,
      "text": "变成一台可以长期放在家里",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1636.123,
      "end": 1638.763,
      "text": "由不同设备随时电用的AI服务器",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1638.763,
      "end": 1641.523,
      "text": "当然这只是一个非常技术的框架",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1641.523,
      "end": 1643.543,
      "text": "后面还可以继续加入更大的模型",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1643.543,
      "end": 1644.383,
      "text": "自己的资料库",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1644.383,
      "end": 1645.323,
      "text": "手机快据指令",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1645.323,
      "end": 1646.923,
      "text": "甚至更多自动化流程",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1646.923,
      "end": 1649.223,
      "text": "大家完全可以根据自己的Mac配置",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1649.223,
      "end": 1651.023,
      "text": "和实际需求继续往上搭",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1651.023,
      "end": 1653.363,
      "text": "先把这套技术环境稳定跑起来",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1653.363,
      "end": 1654.683,
      "text": "再逐步增加功能",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1654.683,
      "end": 1657.163,
      "text": "更适合第一次接触本地AI的各位",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1657.163,
      "end": 1657.623,
      "text": "好了",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1657.623,
      "end": 1659.603,
      "text": "那么以上就是本期视频的全部内容了",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1659.603,
      "end": 1661.363,
      "text": "如果觉得本期视频有所帮助的话",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1661.363,
      "end": 1663.623,
      "text": "还请不要忘记点赞链加工主支持一下",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1663.623,
      "end": 1664.183,
      "text": "就是李军",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1664.183,
      "end": 1665.543,
      "text": "我们下期视频再见",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    },
    {
      "start": 1665.543,
      "end": 1666.083,
      "text": "拜拜",
      "language": "zh",
      "chunk": 3,
      "local_language_run": 4,
      "language_run": 27,
      "source": "whisper"
    }
  ],
  "transcript_selection": {
    "schema": "videonote.transcript_selection.v1",
    "status": "SELECTED",
    "requested_language": "zh",
    "duration": 1666.751563,
    "reference_path": null,
    "reference_kind": null,
    "reference_language": null,
    "selected_source": "whisper_segmented_multilingual",
    "decision_reason": "multilingual_segmented_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": 731,
      "multilingual": true,
      "first_start": 0.0,
      "final_end": 1666.083,
      "duration_coverage": 0.999599,
      "maximum_gap_seconds": 12.218,
      "inverted_count": 0,
      "language": {
        "requested": "zh",
        "matched": true,
        "latin": 2400,
        "han": 7407,
        "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": 1666.752,
      "rejected_segment_count": 0,
      "rejected_segments": []
    }
  }
}