Path1.Mod1.a - Explore ML Workspace - Setting Up Your Workspace

Question 1

O C R AMLDS AMLCO AMLRU

Six built in RBACs for controlling workspace access

Answer 1

• Owner
• Contributor - can’t grant access
• Reader - read only
• Azure ML Data Scientist - excludes create Compute or edit workspace settings
• Azure ML Compute Operator - can manage all Compute in the workspace
• Azure ML Registry User - Manages a Registry ie read/write/delete/promote Models in the Registry. Cannot create new or delete existing Registries

Question 2

Sequence for creating an ML Service/Resource

The step that creates the ML Workspace

Answer 2

Log into Azure Portal
=> Azure Subscription
==> Create Resource Group
====> Create ML Service/Resource (to create the workspace)
=====> Auto-Created Resources

Question 3

The four Azure Resources created when creating an ML Service/Resource

Answer 3

Azure Storage Account
Azure Key Vault
Azure Insights
Azure Container Registry

Question 4

4 Ways to create a Workspace

Answer 4

1. Azure Portal
2. ARM Template
3. Azure CLI
4. Azure ML Python SDK

Question 5

Study Code example; Describe what it does:

from azure.ai.ml.entities import Workspace
workspace_name = "mlw-example"

ws_basic = Workspace(
    name=workspace_name,
    location="eastus",
    display_name="Basic workspace-example",
    description="This example shows how to create a basic workspace",
)
ml_client.workspaces.begin_create(ws_basic)

Answer 5

It creates a Workspace named “mlw-example” using the Python SDK, Region “East US”

Question 6

TS E R, DA Dep ReCom

The three organizational setups for ML Solutions

Three common considerations between all three

Answer 6

Team Structure, Environment and Regional Setups

Data Access, Managing Deployment across Environments, Results Comparisons

Question 7

Define Environment and give examples

Answer 7

“A collection of resources targeted by a deployment process, based on their stage in the application lifecycle”

Examples: Dev (Local), Test, QA, Staging, Production