blocks|key|2690811|text|关于训练，您将希望使用导入数据模块。在那里你可以指定如何连接到你的Azure+SQL数据库。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|2690812|这篇文章可能会给你一个很好的pointer，它展示了如何使用Azure+Data+Factory来同步重新训练实验。|offset|length|2690813|entityMap|0|LINK|mutability|MUTABLE|url|https://azure.microsoft.com/en-us/blog/retraining-and-updating-azure-machine-learning-models-with-azure-data-factory/^0|0|E|7|0|0^^$0|@$1|2|3|4|5|6|7|N|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|O|8|@]|9|@$D|P|E|Q|1|R]]|A|$]]|$1|F|3|-4|5|6|7|S|8|@]|9|@]|A|$]]]|G|$H|$5|I|J|K|A|$L|M]]]]

regarding the training, you will want to use the import data modules. In there you can specify how to connect to your Azure SQL database. 

This article might give you a good <a href="https://azure.microsoft.com/en-us/blog/retraining-and-updating-azure-machine-learning-models-with-azure-data-factory/" rel="nofollow noreferrer">pointer</a>, it shows how to use Azure Data Factory to synchronize retraining of an experiment.

blocks|key|24346|text|我发现这很难在文档中找到，但我发现了这个链接，我能够在不上传blob的情况下成功地运行批处理执行。https://docs.microsoft.com/en-us/azure/machine-learning/studio/web-services-that-use-import-export-modules|type|unstyled|depth|inlineStyleRanges|entityRanges|offset|length|data|24347|关键部分是：|24348|24349|将C#示例代码复制并粘贴到Program.cs文件中，删除对blob存储的所有引用|unordered-list-item|24350|找到请求声明并更新Web+Service值|24351|24352|传递给Import+Data和Export+Data的参数|24353|模块。在本例中，您使用原始查询，但定义了一个新的|24354|表名。|24355|要配置导入查询和目标表的Web服务参数，请执行以下操作：|24356|在Import+Data模块的属性窗格中，单击Database+query字段右上角的图标，然后选择Set+as+web+service参数。在Export+Data模块的属性窗格中，单击Data+table+name字段右上角的图标，然后选择Set+as+web+service参数。在“导出数据模块”属性窗格底部的“Web服务参数”部分中，单击“数据库查询”并将其重命名为“查询”。单击Data+Table+name并将其重命名为table。|24357|+var+request+=+new+BatchExecutionRequest()+
+{+++++++++++
+++++GlobalParameters+=+new+Dictionary<string,+string>()+{
+++++++++{+"Query",+@"select+[age],+[workclass],+[fnlwgt],+[education],+[education-num],+[marital-status],+[occupation],+[relationship],+[race],+[sex],+[capital-gain],+[capital-loss],+[hours-per-week],+[native-country],+[income]+from+dbo.censusdata"+},
+++++++++{+"Table",+"dbo.ScoredTable2"+},
+++++}
+};|code-block|syntax|javascript|24358|一旦将数据库用作加载数据模块源，就不需要在训练模块上有web服务输入。您还可以将数据库查询设置为web参数。一旦你运行批处理执行作业来重新训练模型，你就可以将它们存储在Azure+blob存储中，并让你的预测模型在运行时使用“加载训练好的模型”模块而不是训练好的模型模块从那里加载它们。有关该过程，请参阅此链接：|24359|https://blogs.technet.microsoft.com/machinelearning/2017/06/19/loading-a-trained-model-dynamically-in-an-azure-ml-web-service/|24360|所以简而言之：|24361|将SQL数据库用作导入数据模块的源|24362|按所需的时间间隔运行批处理过程以重新训练模型|24363|将重新训练的模型(ilearner文件)保存到blob存储中，或保存到可访问的http地址|24364|在预测实验中使用负载训练模型模块，而不是训练模型。|24365|将您的blob或url的路径放入load+trained模型模块的参数中。|24366|使用动态加载的模型运行、发布和测试预测实验|24367|请注意，如果您的实验中有多个模型，则可以使用此方法。|24368|entityMap|0|LINK|mutability|MUTABLE|url|https://docs.microsoft.com/en-us/azure/machine-learning/studio/web-services-that-use-import-export-modules|1^0|1D|2Y|0|0|0|0|0|0|0|0|0|0|0|0|0|0|0|3I|1|0|0|0|0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|1U|8|@]|9|@$A|1V|B|1W|1|1X]]|C|$]]|$1|D|3|E|5|6|7|1Y|8|@]|9|@]|C|$]]|$1|F|3|-4|5|6|7|1Z|8|@]|9|@]|C|$]]|$1|G|3|H|5|I|7|20|8|@]|9|@]|C|$]]|$1|J|3|K|5|I|7|21|8|@]|9|@]|C|$]]|$1|L|3|-4|5|6|7|22|8|@]|9|@]|C|$]]|$1|M|3|N|5|6|7|23|8|@]|9|@]|C|$]]|$1|O|3|P|5|6|7|24|8|@]|9|@]|C|$]]|$1|Q|3|R|5|6|7|25|8|@]|9|@]|C|$]]|$1|S|3|T|5|6|7|26|8|@]|9|@]|C|$]]|$1|U|3|V|5|6|7|27|8|@]|9|@]|C|$]]|$1|W|3|X|5|Y|7|28|8|@]|9|@]|C|$Z|10]]|$1|11|3|12|5|6|7|29|8|@]|9|@]|C|$]]|$1|13|3|14|5|6|7|2A|8|@]|9|@$A|2B|B|2C|1|2D]]|C|$]]|$1|15|3|16|5|6|7|2E|8|@]|9|@]|C|$]]|$1|17|3|18|5|6|7|2F|8|@]|9|@]|C|$]]|$1|19|3|1A|5|6|7|2G|8|@]|9|@]|C|$]]|$1|1B|3|1C|5|6|7|2H|8|@]|9|@]|C|$]]|$1|1D|3|1E|5|6|7|2I|8|@]|9|@]|C|$]]|$1|1F|3|1G|5|6|7|2J|8|@]|9|@]|C|$]]|$1|1H|3|1I|5|6|7|2K|8|@]|9|@]|C|$]]|$1|1J|3|1K|5|6|7|2L|8|@]|9|@]|C|$]]|$1|1L|3|-4|5|6|7|2M|8|@]|9|@]|C|$]]]|1M|$1N|$5|1O|1P|1Q|C|$1R|1S]]|1T|$5|1O|1P|1Q|C|$1R|14]]]]

I found this difficult to find in the documentation, but found this link that I was able to successfully run a batch execution without uploading a blob. <a href="https://docs.microsoft.com/en-us/azure/machine-learning/studio/web-services-that-use-import-export-modules" rel="nofollow noreferrer">https://docs.microsoft.com/en-us/azure/machine-learning/studio/web-services-that-use-import-export-modules</a>

The key part is:

<ul>
<li>Copy and paste the C# sample code into your Program.cs file, and
remove all references to the blob storage</li>
<li>Locate the request declaration and update the values of Web Service 
Parameters that are passed to the Import Data and Export Data 
modules. In this case, you use the original query, but define a new 
table name.</li>
</ul>

To configure the Web Service Parameters for the import query and the destination table:

In the properties pane for the Import Data module, click the icon at the top right of the Database query field and select Set as web service parameter.
In the properties pane for the Export Data module, click the icon at the top right of the Data table name field and select Set as web service parameter.
At the bottom of the Export Data module properties pane, in the Web Service Parameters section, click Database query and rename it Query.
Click Data table name and rename it Table.

<pre><code> var request = new BatchExecutionRequest() 
 { 
 GlobalParameters = new Dictionary&lt;string, string&gt;() {
 { "Query", @"select [age], [workclass], [fnlwgt], [education], [education-num], [marital-status], [occupation], [relationship], [race], [sex], [capital-gain], [capital-loss], [hours-per-week], [native-country], [income] from dbo.censusdata" },
 { "Table", "dbo.ScoredTable2" },
 }
 };
</code></pre>

Once you have the database as your load data module source, you need not have a web service input on the training module. You can also set a database query as a web parameter. Once you run the batch execution job to retrain the models, you can store them in Azure blob storage, and have your predictive model load them from there at runtime using "load trained model" modules rather than trained model modules. See this link for that procedure:

<a href="https://blogs.technet.microsoft.com/machinelearning/2017/06/19/loading-a-trained-model-dynamically-in-an-azure-ml-web-service/" rel="nofollow noreferrer">https://blogs.technet.microsoft.com/machinelearning/2017/06/19/loading-a-trained-model-dynamically-in-an-azure-ml-web-service/</a>

So in short:

Use your SQL database as the source for your import data module

Run the batch execution process to retrain the model at whatever interval you want

Save the retrained models (ilearner files) to blob storage, or at an http address that is accessable

Use the load trained model module in your predictive experiment rather than trained model.

put the path to your blob or url in the parameters of the load trained model module(s). 

run, publish and test the predictive experiment with the dynamically loaded models

Note, this approach can be used if you have multiple models in your experiment.

I am new in Azure ML Studio and try to figure out how to use Azure SQL Server as training input parameter.
So far I could sync local data into Azure SQL database and train data and get output on Azure ML Studio.
I've deployed prediction web service as well.
I need to trigger training service to read data from Azure SQL server but all examples around show Azure blob storage as training service input.

<ol>
<li>Is there a way to use Azure SQL Server as training input parameter and whole training data will be read from that database and the trained model get updated ?</li>
<li>Do I need save trained model to Azure Blob storage to be used in predictive service or it will be updated after training batch job completed and will predictive service start using trained model automatically ?</li>
</ol>

Using Azure SQL Server as Azure ML Training Service Input Parameter

数据同步

SQL数据库

SQL Server

数据库

服务器

我是Azure ML Studio的新手，正在尝试弄清楚如何使用Azure SQL Server作为训练输入参数。到目前为止，我可以将本地数据同步到Azure SQL数据库中，并在Azure ML Studio上训练数据和获得输出。我也部署了预测web服务。我需要触发训练服务来从Azure SQL服务器读取数据，但周围...

问将Azure SQL Server用作Azure ML Training Service输入参数
EN

社区

活动

资源

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问将Azure SQL Server用作Azure ML Training Service输入参数EN

社区

活动

资源

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问将Azure SQL Server用作Azure ML Training Service输入参数
EN