In this post series, I am going to show how we can use Azure Machine learning services and the new features added that make life so easy to train, deploy, automate managing machine learning models . In this post, first I will show how to use a no code environment for Auto ML, how to access it and some difference between Azure mL Studio and services.
Azure ML Service VS Azure ML Studio
Before going ahead, there is a question of what is the difference between Azure Machine learning Studio and Services
Azure ML Studio
is a drag and drops environment and No Coding required. There are some Pre Build Algorithms and data transformation tools.
When to use it: experiment ML models easily, built-in algorithms able to address your issues
Azure ML Services
it is a coding environment, you can access to most important open source machine learning frameworks such as Pytorch, Tensorflow and so forth. It is an environment for Python coding, and you have control over your ML algorithms or any free libraries.
in this and next post I will start with below topics
1- How to set up the environment
2- Image classification with MNIST Database
3- Regression model
4- Automated Machine learning
6- Machine learning Pipelines
7- Create and Consume Webservice in Power BI Services and Desktop
8- and many more discussions
Access to Azure Machine Learning Services
The first step is to create a workspace,
what is workspace: it is the main place that we put all the resources we created there.
I will create a workspace in Azure Portal to do that, you need to have a valid account for Azure if you do not have, follow the below link
you will have access to a 12-month free account
if you already have an Azure Portal, then you need to create Machine Learning Services Workspace
Next, after creating the machine learning services, you will see a component has been created. As you can see in the below picture
Number 2- we have the main workspace to explore the Azure ML Services
Number 3– Access to Azure ML Service forum
Number 4- some free samples posted in Github
Number 5- we access to some documentation
Number 6-Build a model using the Visual Interface (preview)
Number 7- Create Notebooks in Azure ML Notebook VM (Preview)
Number 8- Create a new Automated Machine Learning Model (Preview)
in this scenario, click on the Create a New Automated Machine Learning Model (Preview).
This is AutoML, what is mean it just get the data, the type of the problem you want to solve and configure where you want to compute then deploy it and you able to consume it in other applications.
Now Click on the Create a New Automated Machine Learning Model (Preview).
Once you created, the main overview of the completes, running, failed, and other Aautoml practice will be shown. As you can see in the above picture, some information about what day they run, the number of the completed model and so forth have been shown. Also you able to see five different pipelines in the Assets
for the next post, I will show how to create an Experiment and see the run it for the Titanic example!