Provisioning compute resources
Compute resources allow you to execute code scripts during your data exploratory analysis, the training phase, and when operationalizing ML models. The Azure ML workspace offers the following types of compute resources:
- Compute instances: These are virtual machines dedicated to each data scientist that is working in the Azure ML workspace.
- Compute clusters: These are scalable computer clusters that can run multiple training or inference steps in parallel.
- Inference clusters: These are Azure Kubernetes Service (AKS) clusters that can operationalize Docker images, which expose your models through a REST API.
- Attached compute: These are existing compute resources, such as Ubuntu Virtual Machines (VMs) or Synapse Spark pools, that can be attached to the workspace to execute some of the steps of your training or inference pipelines.
When you visit the Manage | Compute section of Azure ML Studio, you will see and be able to manage...
