{"id":508,"date":"2021-08-27T16:20:21","date_gmt":"2021-08-27T08:20:21","guid":{"rendered":"https:\/\/aif.amtbbs.org\/?p=508"},"modified":"2021-08-27T16:20:21","modified_gmt":"2021-08-27T08:20:21","slug":"%e4%ba%9a%e9%a9%ac%e9%80%8a%e6%ad%a3%e5%9c%a8%e9%87%8d%e5%a1%91mlops","status":"publish","type":"post","link":"https:\/\/aif.amtbbs.org\/index.php\/2021\/08\/27\/508\/","title":{"rendered":"\u4e9a\u9a6c\u900a\u6b63\u5728\u91cd\u5851MLOps"},"content":{"rendered":"
\"\u56fe\u7247\"<\/section>\n
\u4f5c\u8005 | Vishnu Prathish<\/section>\n
\u8bd1\u8005 | \u738b\u5f3a<\/section>\n
\u7b56\u5212 | \u51ac\u6885<\/section>\n

\u672c\u6587\u6700\u521d\u53d1\u5e03\u4e8e Medium \u7f51\u7ad9\uff0c\u7ecf\u539f\u4f5c\u8005\u6388\u6743\u7531 InfoQ \u4e2d\u6587\u7ad9\u7ffb\u8bd1\u5e76\u5206\u4eab\u3002<\/strong><\/em><\/p>\n

\u4f17\u6240\u5468\u77e5\uff0c\u5728\u4e09\u5927\u4e91\u63d0\u4f9b\u5546\u4e2d AWS \u62e5\u6709\u6700\u4e30\u5bcc\u7684\u673a\u5668\u5b66\u4e60\u80fd\u529b\u7ec4\u5408\u3002\u968f\u7740 Sagemaker Studio \u4e8e 2020 \u5e74\u521d\u516c\u5f00\u53d1\u5e03\uff0c\u4ed6\u4eec\u521b\u5efa\u4e86\u4e00\u4e2a\u5168\u96c6\u6210\u7684 ML \u5f00\u53d1\u73af\u5883\u2014\u2014\u8fd9\u662f\u4e1a\u754c\u9996\u521b\u3002<\/p>\n

\u5728\u6240\u6709 ML \u4ea7\u54c1\u7684\u4e2d\u5fc3\u951a\u5b9a\u4e00\u4e2a IDE \u662f\u4e00\u4e2a\u660e\u667a\u7684\u4e3e\u52a8\u2014\u2014\u53ea\u8981\u4f60\u7684\u76f8\u5173\u670d\u52a1\u6b63\u786e\u5730\u586b\u8865\u4e86\u5173\u952e\u8fd0\u7ef4\u5c42\u9762\u7684\u7a7a\u767d\u3002\u5982\u679c\u4e00\u5207\u987a\u5229\uff0c\u4e9a\u9a6c\u900a\u5c06\u6709\u673a\u4f1a\u4e00\u52b3\u6c38\u9038\u5730\u91cd\u5851\u884c\u4e1a\u4e2d\u673a\u5668\u5b66\u4e60\u7684\u9762\u8c8c\u3002<\/p>\n

\u751a\u81f3\u5728 Sagemaker Studio \u4e4b\u524d\uff0cAWS \u5c31\u6709\u4e86\u4e00\u4e9b\u9488\u5bf9 MLOps \u7684\u670d\u52a1\u3002\u4f46\u662f\uff0cRe:invent 2020 \u66f4\u8fdb\u4e00\u6b65\u3002\u4ed6\u4eec\u53d1\u5e03\u4e86\u4e00\u7cfb\u5217\u4ea7\u54c1 \/ \u670d\u52a1\uff0c\u586b\u8865\u4e86\u5927\u591a\u6570\u5df2\u77e5\u7684\u7a7a\u767d\u3002<\/p>\n

\u73b0\u5728\u4ed6\u4eec\u505a\u5f97\u600e\u4e48\u6837\uff1f\u4ed6\u4eec\u662f\u5426\u4e3a\u6b63\u786e\u7684\u53d7\u4f17\u6784\u5efa\u4e86\u6b63\u786e\u7684\u5de5\u5177\uff1f\u8fd9\u4e2a\u95ee\u9898\u8fd8\u9700\u8981\u51e0\u5e74\u65f6\u95f4\u624d\u80fd\u5f97\u51fa\u7b54\u6848\u3002\u4f46\u662f AWS \u80af\u5b9a\u5728\u8fd9\u573a\u7ade\u8d5b\u4e2d\u5904\u4e8e\u9886\u5148\u4f4d\u7f6e\u3002\u73b0\u5728\uff0c\u6211\u4eec\u6765\u7814\u7a76\u4e00\u4e9b\u5173\u952e\u7684\u65b0\u670d\u52a1\uff0c\u4ece\u4e2d\u4e86\u89e3 AWS \u5728\u8fd9\u573a\u6e38\u620f\u4e2d\u7684\u4f18\u52bf\u6240\u5728\u3002<\/p>\n

AWS \u7684\u73b0\u6709 MLOps \u5957\u4ef6<\/section>\n

\u4e9a\u9a6c\u900a\u7684\u73b0\u6709\u4ea7\u54c1\u5b8c\u5168\u57fa\u4e8e Sagemaker Studio\u3002\u5b83\u4e3a ML \u5f00\u53d1\u63d0\u4f9b\u4e86\u4e1a\u5185\u9996\u521b\u7684\u96c6\u6210\u5f00\u53d1\u73af\u5883\u3002\u4e0b\u9762\u4ecb\u7ecd\u4e00\u4e9b\u57fa\u4e8e\u5b83\u5b9e\u73b0\u7684\u529f\u80fd\uff0c\u8fd9\u4e9b\u529f\u80fd\u8ba9\u8fd9\u4e2a\u5e73\u53f0\u9887\u5177\u5438\u5f15\u529b\uff1a<\/p>\n

    \n
  • \n
    Sagemaker Studio notebooks \u63d0\u4f9b\u65e0\u670d\u52a1\u5668\u7684 Jupyter \u7b14\u8bb0\u672c\u4ee3\u66ff\u4f60\u7684\u672c\u5730\u7b14\u8bb0\u672c\u3002\u5b83\u8fd8\u652f\u6301\u672c\u5730\u6a21\u5f0f\u3002\u4f46\u6211\u5f3a\u70c8\u5efa\u8bae\u4f60\u56f4\u7ed5\u4e2d\u5fc3\u5316\u7b14\u8bb0\u672c\u8bbe\u7f6e\u6784\u5efa\u5f00\u53d1\u73af\u5883\u3002<\/section>\n<\/li>\n
  • \n
    Sagemaker Autopilot \u5c06 AutoML \u5f15\u5165\u4e86 AWS\uff0c\u4ece\u800c\u6d88\u9664\u4e86 ML \u6d41\u7a0b\u4e2d\u7684\u6240\u6709\u7e41\u91cd\u5de5\u4f5c\u3002<\/section>\n<\/li>\n
  • \n
    Sagemaker Experiments \u5141\u8bb8\u4f60\u4fdd\u5b58\u548c\u8ddf\u8e2a\u4f60\u7684\u8bad\u7ec3\u5b9e\u9a8c\u3002\u5b83\u8fd8\u5141\u8bb8\u5c06\u4e00\u4e2a\u6a21\u578b\u4e0e\u53e6\u4e00\u4e2a\u6a21\u578b\u5bf9\u6bd4\uff0c\u4ece\u800c\u5141\u8bb8\u7528\u6237\u4ece\u5b9e\u9a8c\u7ed3\u679c\u8868\u4e2d\u624b\u52a8\u9009\u62e9\u6700\u4f73\u6a21\u578b\u3002<\/section>\n<\/li>\n
  • \n
    Sagemaker Model tuning \u5141\u8bb8\u4f60\u5229\u7528\u4e91\u6765\u81ea\u52a8\u6267\u884c\u8d85\u53c2\u6570\u4f18\u5316\u3002<\/section>\n<\/li>\n
  • \n
    Multimodel endpoints \u80fd\u5927\u5927\u964d\u4f4e\u63a8\u7406\u6210\u672c\u3002<\/section>\n<\/li>\n
  • \n
    Model monitor \u80fd\u5e2e\u52a9\u4f60\u8ddf\u8e2a\u751f\u4ea7\u4e2d\u7684\u6307\u6807\uff0c\u4ece\u800c\u8f7b\u677e\u8ddf\u8e2a\u6a21\u578b\u6f02\u79fb\u3002<\/section>\n
    2021 \u5e74\u6709\u4ec0\u4e48\u65b0\u53d8\u5316\uff1f<\/section>\n<\/li>\n<\/ul>\n

    \u5c3d\u7ba1 AWS \u662f ML \u670d\u52a1\u7684\u8fd0\u7ef4\u63d0\u4f9b\u5546\uff0c\u4f46\u5b83\u4ecd\u7136\u4e0d\u80fd\u58f0\u79f0\u81ea\u5df1\u62e5\u6709\u7528\u4e8e\u6240\u6709\u673a\u5668\u5b66\u4e60\u76ee\u7684\u7684\uff0c\u6253\u901a\u7684\u5f00\u53d1\u73af\u5883\u3002MLOps \u5728\u51e0\u4e2a\u9886\u57df\u5b58\u5728\u91cd\u5927\u5dee\u8ddd\u3002<\/p>\n

    \u6ca1\u6709\u8fde\u8d2f\u7684 CI\/CD \u7ba1\u9053\u53ef\u4ee5\u5c06\u5b83\u4eec\u8fde\u5728\u4e00\u8d77\u3002\u6ca1\u6709\u8fd9\u6837\u7684\u7ba1\u9053\uff0c\u611f\u89c9\u7528\u6237\u5728\u4f7f\u7528\u4e00\u7cfb\u5217\u4e0d\u540c\u7684\u670d\u52a1\u3002\u673a\u5668\u5b66\u4e60\u8fc7\u7a0b\u5404\u4e2a\u9636\u6bb5\uff08\u6570\u636e\u51c6\u5907\u3001\u8bad\u7ec3\u3001\u9a8c\u8bc1\u3001\u63a8\u7406\u3001\u76d1\u63a7\uff09\u7684\u76f8\u5173\u4ea7\u54c1\u4e5f\u8fd8\u4e0d\u5b8c\u6574\u3002<\/p><\/blockquote>\n

    \u4f46\u8fd9\u79cd\u60c5\u51b5\u6b63\u5728\u6539\u53d8\u3002\u968f\u7740 re:invent 2020 \u548c\u4e4b\u524d\u53d1\u5e03\u7684\u4e00\u4e9b\u65b0\u670d\u52a1\u7684\u51fa\u73b0\uff0cAWS \u5728\u4eca\u5e74\u5df2\u7ecf\u586b\u8865\u4e86\u5927\u90e8\u5206\u7a7a\u767d\uff0c\u800c\u5176\u4ed6\u591a\u6570\u63d0\u4f9b\u5546\u5219\u8fdc\u8fdc\u843d\u540e\u3002<\/p>\n

    \u4e0b\u9762\u662f\u4e00\u4e9b\u4f8b\u5b50\u3002<\/p>\n

    \n
    \n
    Data Wrangler\uff1a\u96f6\u4ee3\u7801\u6570\u636e\u51c6\u5907<\/section>\n<\/section>\n<\/section>\n

    AWS Sagemaker Data Wrangler \u63d0\u4f9b\u4e86\u4e00\u79cd\u5e72\u51c0\u7684 Jupyter \u98ce\u683c\u7684 IDE\uff0c\u7528\u4e8e\u673a\u5668\u5b66\u4e60\u6570\u636e\u51c6\u5907\u3002\u5b83\u76f4\u63a5\u5efa\u7acb\u5728 Sagemaker Studio \u4e0a\uff0c\u56e0\u6b64\u5229\u7528\u4e86 Studio \u7684\u6240\u6709\u5f3a\u5927\u529f\u80fd\uff08\u6bd4\u5982\u5b83\u7684\u6570\u636e\u53ef\u89c6\u5316\uff09\u3002<\/p>\n

    \u5373\u4f7f\u4ece\u6280\u672f\u4e0a\u8bb2\u8fd9\u662f\u4e00\u79cd\u65e0\u4ee3\u7801\u5de5\u5177\uff0c\u4f46 Data Wrangler \u8fd8\u662f\u53ef\u4ee5\u4f7f\u7528\u4ee3\u7801\u81ea\u5b9a\u4e49\u7684\u3002\u4f60\u53ef\u4ee5\u5c06 300 \u591a\u79cd\u5185\u7f6e\u7684\u81ea\u52a8\u8f6c\u6362\u5e94\u7528\u4e8e\u4f60\u7684\u8bad\u7ec3\u6570\u636e\u3002\u4f60\u53ea\u9700\u5355\u51fb\u5373\u53ef\u5c06\u5de5\u4f5c\u6d41\u7a0b\u5bfc\u51fa\u5230 Sagemaker \u7b14\u8bb0\u672c\u5e76\u6784\u5efa\u5c31\u5730\u6a21\u578b\u3002\u5b83\u8fd8\u76f4\u63a5\u652f\u6301\u591a\u4e2a\u6570\u636e\u5b58\u50a8\uff0c\u5305\u62ec Snowflake\u3001MongoDB \u548c Databricks\u3002<\/p>\n

    Data Wrangler \u89e3\u51b3\u4e86\u4e9a\u9a6c\u900a\u5728 ML \u6570\u636e\u51c6\u5907\u65b9\u9762\u7684\u5de8\u5927\u7a7a\u767d\u3002\u4ed6\u4eec\u58f0\u79f0\uff0c\u4ee5\u8fd9\u79cd\u65b9\u5f0f\u7b80\u5316\u6570\u636e\u51c6\u5907\u5de5\u4f5c\u53ef\u4ee5\u5927\u5927\u51cf\u5c11\u7528\u6237\u82b1\u8d39\u5728\u6570\u636e\u51c6\u5907\u4e0a\u7684\u65f6\u95f4\u3002<\/p>\n

    Sagemaker DataBrew\uff1a\u540c\u6837\u7684\u5de5\u5177\uff0c\u4f46\u505a\u6cd5\u4e0d\u540c<\/section>\n

    \"\u56fe\u7247\"<\/p>\n

    Data Brew \u4e5f\u662f\u65e0\u4ee3\u7801\u6570\u636e\u51c6\u5907\u5de5\u5177\u3002\u4f46\u8fd9\u4e24\u79cd\u5de5\u5177\u9762\u5411\u7684\u662f\u4e24\u7c7b\u4e0d\u540c\u7684\u53d7\u4f17\u3002Data Wrangler \u4e13\u95e8\u9488\u5bf9 ML\uff0c\u800c Data Brew \u4e13\u6ce8\u5728\u901a\u7528\u63a2\u7d22\u6027\u6570\u636e\u5206\u6790\uff08EDA\uff09\u4e0a\u3002\u53e6\u5916\uff0cData Brew \u662f\u4e00\u4e2a\u4ee5 UI \u4e3a\u4e2d\u5fc3\u7684\u5de5\u5177\u3002<\/p>\n

    EDA \u901a\u5e38\u662f ML \u7684\u5148\u51b3\u6761\u4ef6\uff0c\u56e0\u6b64\u5b83\u4eec\u5b8c\u5168\u53ef\u4ee5\u540c\u65f6\u4f7f\u7528\u3002Data Brew \u7684\u4e00\u952e\u5206\u6790\u548c\u7cbe\u5fc3\u8bbe\u8ba1\u7684\u754c\u9762\uff08\u9002\u5408\u4e0d\u4f1a\u7f16\u5199\u4ee3\u7801\u7684\u7528\u6237\uff09\u8ba9\u4f5c\u4e1a\u53d8\u5f97\u66f4\u52a0\u7b80\u5355\u660e\u4e86\u3002<\/p>\n

    \u4e24\u79cd\u5de5\u5177\u90fd\u53ef\u4ee5\u7528\u6765\u5b8c\u6210\u7279\u5f81\u5de5\u7a0b\u3002\u4f46\u662f\u53ea\u6709 Data Wrangler \u652f\u6301\u5c06\u7279\u5f81\u7a7a\u95f4\u5bfc\u51fa\u5230 AWS Feature Store\uff0c\u6240\u4ee5\u66f4\u5408\u9002\u4e00\u4e9b\u3002<\/p>\n

    \u53e6\u4e00\u4e2a\u7a7a\u767d\uff0c\u586b\u8865\u5b8c\u6bd5\u3002<\/p>\n

    AWS Feature Store\uff1a\u5927\u89c4\u6a21\u7279\u5f81\u5de5\u7a0b<\/section>\n

    \u8fd9\u662f\u4e00\u4e2a\u91cd\u8981\u7684\u53d1\u5e03\uff0c\u89e3\u51b3\u4e86\u5173\u952e\u7684 \u7279\u5f81\u5de5\u7a0b\u7f3a\u5931 \u7684\u95ee\u9898\u3002\u8bb8\u591a\u673a\u5668\u5b66\u4e60\u5b9e\u8df5\u5728\u8131\u673a\uff08\u6279\u5904\u7406\uff09\u548c\u5728\u7ebf\uff08\u5b9e\u65f6\uff09\u7279\u5f81\u5de5\u7a0b\u4e4b\u95f4\u5b58\u5728\u5dee\u5f02\u3002\u590d\u6742\u7684\u7279\u5f81\u5de5\u7a0b\u8f6c\u6362\u548c\u5728\u6279\u5904\u7406\u671f\u95f4\u6784\u5efa\u7684\u65b0\u7279\u5f81\u5f88\u96be\u5f88\u597d\u5730\u8f6c\u6362\u4e3a\u63a8\u7406 \/ \u9884\u6d4b\u7ba1\u9053\u3002<\/p>\n

    Feature Store \u5728\u8fd9\u4e24\u70b9\u4e4b\u95f4\u653e\u7f6e\u4e86\u4e00\u4e2a\u4e13\u6709\u7684\u9488\u5bf9\u7279\u5f81\u7a7a\u95f4\u7684\u5b58\u50a8\u5e93\u6765\u89e3\u51b3\u8fd9\u4e00\u95ee\u9898\u3002\u8bad\u7ec3\u671f\u95f4\u4f60\u5728 Sagemaker Studio \u4e2d\u5bf9\u539f\u59cb\u6570\u636e\u6240\u505a\u7684\u6240\u6709\u64cd\u4f5c\u90fd\u53ef\u4ee5\u5bfc\u51fa\u5230 Feature Store \u4e2d\uff0c\u5e76\u4e14\u53ef\u4ee5\u4fdd\u8bc1\u5728\u63a8\u7406\u8fc7\u7a0b\u4e2d\u53ef\u4ee5\u6b63\u786e\u5730\u590d\u5236\u8fd9\u4e9b\u6570\u636e\u3002<\/p>\n

    \u9664\u4e86\u89e3\u51b3\u8fd9\u4e2a\u5728\u7ebf – \u79bb\u7ebf\u95ee\u9898\u5916\uff0c\u5b83\u8fd8\u652f\u6301\u7279\u5f81\u53ef\u53d1\u73b0\u3001\u5171\u4eab\u548c\u7279\u5f81\u91cd\u7528\u3002\u5b83\u7684\u8bbe\u8ba1\u8fd8\u8003\u8651\u4e86\u5ef6\u8fdf\u2014\u2014\u8fd9\u662f\u5927\u89c4\u6a21\u573a\u666f\u4e2d\u5fc5\u987b\u505a\u7684\u3002<\/p>\n

    Sagemaker Pipelines\uff1a\u673a\u5668\u5b66\u4e60\u6d41\u7a0b\u7684 CI\/CD<\/section>\n

    \u5bf9\u6211\u6765\u8bf4\uff0c\u8fd9\u9879\u670d\u52a1\u662f\u672c\u5e74\u5ea6\u6700\u91cd\u8981\u7684\u8fd0\u7ef4\u53d1\u5e03\u3002<\/p>\n

    \u5c3d\u7ba1\u53ef\u6269\u5c55 ML \u7684\u91cd\u8981\u5148\u51b3\u6761\u4ef6\u662f\u53ef\u9760\u7684 CI\/CD \u6d41\u7a0b \/ \u6846\u67b6\uff0c\u4f46\u4e4b\u524d\u5e76\u6ca1\u6709\u597d\u7528\u7684\u4ea7\u54c1\u9009\u9879\u3002\u5927\u5bb6\u8981\u4e48\u7528\u7684\u662f\u6ca1\u90a3\u4e48\u7406\u60f3\u7684 MLOps \u6d41\u7a0b\uff0c\u8981\u4e48\u5efa\u7acb\u4e86\u81ea\u5df1\u7684 CI\/CD \u7248\u672c\u3002<\/p>\n

    ML \u7684\u81ea\u5236 CI\/CD \u6846\u67b6\u5b58\u5728\u7684\u95ee\u9898\u662f\u5b83\u4eec\u65e0\u6cd5\u63a8\u5e7f\uff0c\u56e0\u6b64\u65e0\u6cd5\u8f7b\u677e\u5f00\u6e90\u3002\u6846\u67b6\u4e0d\u53ef\u907f\u514d\u5730\u8981\u5728\u4ee3\u7801\u4e2d\u5199\u5165\u8bb8\u591a\u9886\u57df\u77e5\u8bc6\u2014\u2014\u65e2\u662f\u4e3a\u4e86\u7f29\u77ed\u5f00\u53d1\u65f6\u95f4\uff0c\u4e5f\u662f\u4e3a\u4e86\u4e0e\u73b0\u6709\u670d\u52a1\u66f4\u597d\u5730\u96c6\u6210\u3002AWS \u6253\u7b97\u901a\u8fc7\u7528\u4e8e ML \u7684\u901a\u7528 CI\/CD \u6846\u67b6\u89e3\u51b3\u8fd9\u4e00\u95ee\u9898\u3002<\/p>\n

    Sagemaker Pipelines \u5141\u8bb8\u4f60\u521b\u5efa\u3001\u53ef\u89c6\u5316\u548c\u7ba1\u7406 ML \u5de5\u4f5c\u6d41\u3002\u5b83\u4f7f\u4f60\u80fd\u591f\u521b\u5efa\u5355\u72ec\u7684\u5f00\u53d1\u548c\u751f\u4ea7\u73af\u5883\u5e76\u8fdb\u884c\u8ddf\u8e2a\u3002\u73af\u5883\u5141\u8bb8\u4f60\u8fdb\u884c\u5de5\u4ef6\u5347\u7ea7\u3002\u5b83\u8fd8\u5e26\u6709\u4e00\u4e2a\u6a21\u578b\u6ce8\u518c\u8868\uff0c\u53ef\u8ba9\u4f60\u8ddf\u8e2a\u548c\u9009\u62e9\u6b63\u786e\u7684\u90e8\u7f72\u6a21\u578b\u3002<\/p>\n

    \u8fd9\u4e00\u7ba1\u9053\u7684\u4e00\u4e2a\u4e0d\u592a\u660e\u663e\u7684\u6548\u679c\u662f\uff0c\u5b83\u8fd8\u5c06\u5176\u4ed6\u6240\u6709\u7528\u4e8e ML \u7684 Sagemaker \u670d\u52a1\u7f16\u7ec7\u5728\u4e00\u8d77\u3002\u8fd9\u4e3a AWS \u5e26\u6765\u4e86\u660e\u663e\u7684\u4f18\u52bf\uff0c\u56e0\u4e3a\u5b83\u53ef\u4ee5\u5b9e\u73b0\u771f\u6b63\u7684\u7aef\u5230\u7aef ML\u3002<\/p>\n

    re:invent \u7684\u5176\u4ed6\u76f8\u5173\u5185\u5bb9<\/section>\n
    Sagemaker Clarify\uff1a<\/section>\n

    \u8de8 e2e Sagemaker \u5de5\u4f5c\u6d41\u7684\u504f\u89c1\u68c0\u6d4b\u3002\u5bf9\u4e8e B2C \u516c\u53f8\u800c\u8a00\u8fd9\u662f\u4e00\u5927\u4f18\u52bf\u3002<\/p>\n

    SageMaker \u8c03\u8bd5\u5668\u7684\u6539\u8fdb<\/section>\n

    \u8bad\u7ec3\u671f\u95f4\u5bf9\u8d44\u6e90\u5229\u7528\u60c5\u51b5\u8fdb\u884c\u76d1\u89c6\u548c\u6df1\u5ea6 profiling\u3002\u7279\u522b\u662f\u5728\u6df1\u5c42\u795e\u7ecf\u7f51\u7edc\u4e0a\u3002<\/p>\n

    \u8fb9\u7f18\u673a\u5668\u5b66\u4e60<\/section>\n

    Sagemaker edge manager \u57fa\u4e8e AWS Neo \u4e4b\u4e0a\uff0c\u5f15\u5165\u4e86\u8fb9\u7f18\u8bbe\u5907\u7684\u6a21\u578b\u7ba1\u7406\u3002\u5982\u679c\u4f60\u5728\u7269\u8054\u7f51\u884c\u4e1a\uff0c\u5b83\u4f1a\u975e\u5e38\u6709\u7528\u3002<\/p>\n

    \u6570\u636e\u5e93 ML \u529f\u80fd<\/section>\n

    \u867d\u7136\u4e0d\u592a\u7b97\u662f MLOps\uff0c\u4f46\u4e9a\u9a6c\u900a\u65b0\u7684\u6570\u636e\u5e93 ML \u670d\u52a1\u786e\u5b9e\u5c5e\u4e8e\u4e00\u4e2a\u5171\u540c\u7684\u4e3b\u9898\u2014\u2014\u5efa\u7acb\u4e00\u4e2a\u5e73\u7a33\u7684\u751f\u4ea7\u7ea7 ML \u6d41\u7a0b\uff0c\u4ece\u800c\u5b8c\u5168\u6d88\u9664\u4e86\u5bf9\u8fd0\u7ef4\u7684\u9700\u6c42\u3002<\/p>\n

      \n
    • Amazon Redshift ML\uff1a\u5c06 Sagemaker Autopilot \u96c6\u6210\u5230 Amazon Redshift \u4e2d<\/li>\n
    • Amazon Neptune ML\uff1a\u96c6\u6210 Graph ml<\/li>\n
    • Amazon Aurora ML\uff1a\u4f7f\u7528 SQL \u67e5\u8be2\u5c06 ML \u76f4\u63a5\u96c6\u6210\u5230 Postgres \u4e2d\u3002<\/li>\n
    • Amazon Athena ML\uff1a\u5728 Athena \u4e0a\u63d0\u4f9b\u7ecf\u8fc7\u9884\u8bad\u7ec3\u7684\u6a21\u578b\u3002
      \n
      \u7ade\u4e89\u5bf9\u624b\u7684\u60c5\u51b5\uff1f<\/section>\n<\/li>\n<\/ul>\n

      Azure Machine Learning \u548c Google Cloud AI platform \u662f\u6392\u540d\u9760\u524d\u7684\u4e91\u63d0\u4f9b\u5546\u4e2d\u7684\u4e24\u5bb6\u5934\u90e8 MlOps \u63d0\u4f9b\u5546\u3002\u4e24\u8005\u90fd\u5177\u6709\u5f3a\u5927\u7684\u7ba1\u9053\u548c CI\/CD \u529f\u80fd\u3002\u4f46\u662f\uff0cGoogle AI \u7ba1\u9053\u4ecd\u5904\u4e8e\u6d4b\u8bd5\u9636\u6bb5\uff0c\u800c\u5176 AWS \u7ade\u54c1\u5df2\u7ecf\u5177\u5907\u901a\u7528\u6027\u3002Azure Machine Learning Studio \u611f\u89c9\u4e0e Sagemaker \u975e\u5e38\u76f8\u4f3c\uff0c\u4f46\u5e76\u6ca1\u6709\u63d0\u4f9b\u90a3\u4e48\u591a\u670d\u52a1\u3002<\/p>\n

      \u5176\u4ed6\u63d0\u4f9b\u5546\u6240\u7528\u7684\u6a21\u5f0f\u5e76\u6ca1\u6709\u5728\u4e2d\u5fc3\u5305\u542b\u96c6\u6210\u7684 IDE\u3002Azure ML Studio \u4f3c\u4e4e\u5728\u8fd9\u65b9\u9762\u505a\u4e86\u5c1d\u8bd5\u3002\u4f46\u5b83\u5728\u529f\u80fd\u96c6\u65b9\u9762\u5c40\u9650\u5f88\u5927\u3002<\/p>\n

      \u4e0e\u5176\u4ed6\u5934\u90e8\u63d0\u4f9b\u5546\u76f8\u6bd4\uff0c\u4e9a\u9a6c\u900a\u786e\u5b9e\u6295\u5165\u4e86\u66f4\u591a\u8d44\u6e90\u6765\u63d0\u4f9b\u66f4\u597d\u7684\u6570\u636e\u79d1\u5b66\u8fd0\u7ef4\u89e3\u51b3\u65b9\u6848\u3002\u8fd9\u6837\u662f\u5426\u53ef\u4ee5\u8ba9\u4ed6\u4eec\u7262\u7262\u5730\u628a\u6301\u6700\u96c6\u6210\u7684 MLOps \u5957\u4ef6\u7684\u9886\u5148\u5730\u4f4d\uff1f\u6211\u60f3\u662f\u8fd9\u6837\u3002\u4e9a\u9a6c\u900a\u5728\u5f00\u53d1\u4e91\u89e3\u51b3\u65b9\u6848\u65b9\u9762\u5177\u6709 3 \u5230 5 \u5e74\u7684\u9886\u5148\u4f18\u52bf\uff08\u6216\u66f4\u591a\uff1f\u8fd9\u91cc\u6211\u627e\u4e0d\u5230\u53c2\u8003\u6570\u636e\uff09\u3002\u4f46\u662f\uff0c\u73b0\u5728\u9884\u6d4b\u8c01\u5c06\u8d62\u5f97 MLOps \u7ade\u8d5b\u8fd8\u4e3a\u65f6\u8fc7\u65e9\u3002<\/p>\n

      \u539f\u6587\u94fe\u63a5\uff1a<\/p>\n

      <\/div>\n

      <\/g><\/svg><\/i> \"Loading\"<\/p>\n

      <\/div>\n","protected":false},"excerpt":{"rendered":"

      \u4f5c\u8005 | Vishnu Prathish \u8bd1\u8005 | \u738b\u5f3a \u7b56\u5212 | \u51ac\u6885 \u672c\u6587\u6700\u521d\u53d1\u5e03\u4e8e Medium \u7f51\u7ad9\uff0c […]<\/p>\n

      <\/div>\n

      <\/g><\/svg><\/i> \"Loading\"<\/p>\n

      <\/div>\n","protected":false},"author":1,"featured_media":509,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[],"class_list":["post-508","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai"],"_links":{"self":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/508","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/comments?post=508"}],"version-history":[{"count":1,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/508\/revisions"}],"predecessor-version":[{"id":511,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/posts\/508\/revisions\/511"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media\/509"}],"wp:attachment":[{"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/media?parent=508"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/categories?post=508"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aif.amtbbs.org\/index.php\/wp-json\/wp\/v2\/tags?post=508"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}