Exploring Automated ML Training

This guide offers a comprehensive exploration of Automated Machine Learning (ML) training, making it ideal for those looking to streamline their ML processes.

Hafeez Baig
20 steps
2 minutes

Azure

This lab is a continuation of the previous one, Please refer to it before proceeding - <https://scribehow.com/viewer/Setting_up_Automated_ML_Training__wdW9LoAQQR2-mjwgnmsVVA>

On the **Automated ML** page, you can see the ML Job is now completed and its duration

Click on the name of the Automated ML which you have created

**Overview** page will open, here you can see its properties

Click on the **Models + child jobs** tab on the top left side

Checkbox the first one which is **VotingEnsemble**

Click on the **Deploy** button and select the option **Real-time endpoint** option

**What is Real-time endpoint in Automated ML?** A **Real-time endpoint** in Automated ML (AutoML) is a deployed web service that allows you to send data and receive **instant predictions** from your trained machine learning model. It is designed for **low-latency, on-demand inference**, making it ideal for scenarios like fraud detection, recommendation engines, or chatbots where immediate responses are required.

On the right side **Deploy** wizard will open, select the **Endpoint** as **New**

The Endpoint name will be pre-populated

The Deployment name will be pre-populated

Scroll the page and click on the **Deploy** button

After sometime you will see **Success: Model deployment is successfully triggered**

Now click on the **Endpoints** option from the left bar

**What are endpoints in Azure AI?** **Endpoints in Azure AI** are secure URLs through which AI services or models can be accessed and used by client applications. They allow external systems to send data and receive responses from models for tasks like prediction, classification, or content generation. There are typically two types: - **Real-time endpoints**: Provide immediate, on-demand inference. - **Batch endpoints**: Process large volumes of data asynchronously. Endpoints manage authentication, traffic, and deployment versions, enabling seamless integration of AI capabilities into applications.

**Endpoints** page will open, click on the name of the end point which we have created

**Endpoint** details page will open, Click on the **Refresh** button on the top left side

Click on the **Test** tab on the top left side **Note:** The Deployment will take some time, around 45 minutes to 1 hour. If you cannot see the Test tab, then the deployment is not yet completed.

**Test** page will open, here you can see the Input sample interface and test it

Click on the **Consume** tab, here you can view the **REST endpoints** and **Authentication keys**

Scroll the page to the Consumption option, here you can view the sample Code in different languages

You can click on **Python** tab to view the sample code in python

Click on the **Logs** tab, here you can view all the deployment logs

Congratulations on completing this lab and exploring Automated ML Training!

Refer to this lab to continue learning - <https://scribehow.com/viewer/Creating_Azure_Machine_Learning_Pipeline__fx5Ai4fzQbypi3O7sqxsig>