ML Builder is currently in a public early access program (EAP) and you need to register at the ML Builder registration page to try it out.
To get started with ML Builder, do the following:
Sign in to the ML Builder portal and register your non-production environment. When signing in, enter the name of your environment in Environment. For example, if the URL of the environment is https://pp.example.com enter pp.example.com. This is the source environment ML Builder connects to get data for training models.
Click on How to select the best environment? to read more about which environment to use.
Configure ML Builder to work with your Microsoft Azure subscription and set up your training machines and workspaces.
Install the ML Builder plugin in your development environment. In one of the setup steps click Download the ML Builder plugin to download the plugin and install it in the development environment where you develop apps and try out your models.
Create a model in the ML Builder app. Select your use case depending on what you want to predict, choose data, and let ML Builder create a model.
Validate and deploy your model in the overview page, to make the model available in your environment.
Implement the model predictions in your app.
ML Builder requires Azure to create a model. The first time you sign in, ML Builder asks you to set up your environment to work with Azure. Follow the instructions in ML Builder to configure the settings.
You need the following information from Azure:
- Subscription ID
- Application / Client ID
- Directory / Tenant ID
- Client secret
- Machine learning workspace and a training cluster
- Storage container and the connection string for storing data
Here is an overview of how ML Builder uses Azure.