Hello, in this article we will see how to use the Azure ML service, we will start the creation of a workspace and then we will see the creation and execution of an experiment.
- A Microsoft Azure subscription.
Disclaimer: Given the constant updating of Cloud Computing technologies some steps may be different when reading the article, I will make the effort to keep it updated, but there could be some differences between what is shown below and the Azure console at the time of the implementation.
Creation of a Workspace
- Enter the Azure portal at https://portal.azure.com and click on Create a resource
- In the new window click on AI + Machine Learning and then click on Machine Learning service workspace
- Enter the following values:
- Workspace Name: ws-vvillar
- Subscription: <Your subscription>
- Resource group: rg-vvillar
- Location: East US 2
- Once the resource is created click on Go to resource
Creation of an Experiment
- Now that we have created the workspace, we will create an experiment, to do this click on Experiments and then click on Open Azure Notebooks.
- In the new tab click on clone and then open the first notebook called 01.run-experiment.ipynb
- In the notebook go to the first cell and click on Run located at the top of the windows. This will import the necessary packages to create the experiment. (If a red warning message appears ignore it).
- Go to the next cell and click on Run, a URL and a code will appear.
- Open the link and enter the code to authenticate.
- Once authenticated, return to the notebook and click on Run in the next cell. This will create the experiment called my-first-experiment within the workspace.
- In the next cell click on Run. This will make an estimate of the PI value and the run history will be saved in a file.
- Return to the Azure portal you’ll see that you have the experiment created.
- Return to the notebook, in the next cell when clicking on Run you’ll execute the run function that allow you to see the execution of the experiment inside the notebook.
- If you click on Link to Azure Portal it will take you to the details of the result of the experiment.
- If you execute the experiment function you’ll see the experiment inside the notebook.
- By clicking on Link to Azure Portal you’ll see the details of the experiment execution.