[{"data":1,"prerenderedAt":179},["ShallowReactive",2],{"article-other\u002Fmlops_rabbitmq":3},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":7,"title":8,"description":9,"date":10,"tags":11,"body":13,"_type":173,"_id":174,"_source":175,"_file":176,"_stem":177,"_extension":178},"\u002Farticles\u002Fother\u002Fmlops_rabbitmq","other",false,"","MLOps 与 RabbitMQ：模型生命周期与异步任务传递","用 AI 制药平台的分子预测任务说明 MLOps 和 RabbitMQ 的分工：前者管理模型从开发到监控的生命周期，后者负责服务间可靠的异步消息传递。","2026-08-14",[12],"随笔",{"type":14,"children":15,"toc":171},"root",[16,74,100,106,137,142,155,165],{"type":17,"tag":18,"props":19,"children":23},"element","pre",{"className":20,"code":21,"language":22,"meta":7,"style":7},"language-text shiki shiki-themes github-dark","机器学习模型\n   ↓\nMLOps：管理模型从开发、部署到监控的整个生命周期\n   ↓\nRabbitMQ：在不同服务之间可靠地传递任务和消息\n","text",[24],{"type":17,"tag":25,"props":26,"children":27},"code",{"__ignoreMap":7},[28,39,48,57,65],{"type":17,"tag":29,"props":30,"children":33},"span",{"class":31,"line":32},"line",1,[34],{"type":17,"tag":29,"props":35,"children":36},{},[37],{"type":22,"value":38},"机器学习模型\n",{"type":17,"tag":29,"props":40,"children":42},{"class":31,"line":41},2,[43],{"type":17,"tag":29,"props":44,"children":45},{},[46],{"type":22,"value":47},"   ↓\n",{"type":17,"tag":29,"props":49,"children":51},{"class":31,"line":50},3,[52],{"type":17,"tag":29,"props":53,"children":54},{},[55],{"type":22,"value":56},"MLOps：管理模型从开发、部署到监控的整个生命周期\n",{"type":17,"tag":29,"props":58,"children":60},{"class":31,"line":59},4,[61],{"type":17,"tag":29,"props":62,"children":63},{},[64],{"type":22,"value":47},{"type":17,"tag":29,"props":66,"children":68},{"class":31,"line":67},5,[69],{"type":17,"tag":29,"props":70,"children":71},{},[72],{"type":22,"value":73},"RabbitMQ：在不同服务之间可靠地传递任务和消息\n",{"type":17,"tag":75,"props":76,"children":77},"ul",{},[78,90],{"type":17,"tag":79,"props":80,"children":81},"li",{},[82,88],{"type":17,"tag":83,"props":84,"children":85},"strong",{},[86],{"type":22,"value":87},"MLOps",{"type":22,"value":89},"：Machine Learning + Operations。可以理解为“机器学习领域的 DevOps”，负责数据版本、模型训练、测试、部署、监控和更新。它是一套工程方法，不是某个具体软件。",{"type":17,"tag":79,"props":91,"children":92},{},[93,98],{"type":17,"tag":83,"props":94,"children":95},{},[96],{"type":22,"value":97},"RabbitMQ",{"type":22,"value":99},"：一种消息队列中间件。生产者把消息放入队列，消费者异步取出处理，双方不必同时在线。",{"type":17,"tag":101,"props":102,"children":103},"p",{},[104],{"type":22,"value":105},"例如 AI 制药平台提交一个分子预测任务：",{"type":17,"tag":18,"props":107,"children":109},{"className":20,"code":108,"language":22,"meta":7,"style":7},"Java 后端 → RabbitMQ 队列 → Python 模型服务\n                              ↓\n                         执行预测并返回结果\n",[110],{"type":17,"tag":25,"props":111,"children":112},{"__ignoreMap":7},[113,121,129],{"type":17,"tag":29,"props":114,"children":115},{"class":31,"line":32},[116],{"type":17,"tag":29,"props":117,"children":118},{},[119],{"type":22,"value":120},"Java 后端 → RabbitMQ 队列 → Python 模型服务\n",{"type":17,"tag":29,"props":122,"children":123},{"class":31,"line":41},[124],{"type":17,"tag":29,"props":125,"children":126},{},[127],{"type":22,"value":128},"                              ↓\n",{"type":17,"tag":29,"props":130,"children":131},{"class":31,"line":50},[132],{"type":17,"tag":29,"props":133,"children":134},{},[135],{"type":22,"value":136},"                         执行预测并返回结果\n",{"type":17,"tag":101,"props":138,"children":139},{},[140],{"type":22,"value":141},"这里：",{"type":17,"tag":75,"props":143,"children":144},{},[145,150],{"type":17,"tag":79,"props":146,"children":147},{},[148],{"type":22,"value":149},"RabbitMQ 负责排队、削峰、失败重试和服务解耦。",{"type":17,"tag":79,"props":151,"children":152},{},[153],{"type":22,"value":154},"MLOps 负责模型采用哪个版本、如何部署、预测质量是否下降、何时重新训练。",{"type":17,"tag":101,"props":156,"children":157},{},[158,160],{"type":22,"value":159},"一句话：",{"type":17,"tag":83,"props":161,"children":162},{},[163],{"type":22,"value":164},"MLOps 管理模型的生命周期，RabbitMQ 负责系统之间的任务传递。",{"type":17,"tag":166,"props":167,"children":168},"style",{},[169],{"type":22,"value":170},"html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}",{"title":7,"searchDepth":41,"depth":41,"links":172},[],"markdown","content:articles:other:MLOps_RabbitMQ.md","content","articles\u002Fother\u002FMLOps_RabbitMQ.md","articles\u002Fother\u002FMLOps_RabbitMQ","md",1789465115003]