Azure Machine Learning Documentation

Getting Started with Azure Machine Learning

Welcome to Azure Machine Learning! This guide will walk you through the essential steps to begin your journey in creating and managing machine learning solutions on Azure. Whether you're a seasoned data scientist or just starting, Azure ML offers a powerful and flexible platform to accelerate your AI development.

Prerequisites

Before you begin, ensure you have the following:

• An Azure Subscription: If you don't have one, you can create a free account.
• Basic understanding of Machine Learning concepts.
• (Optional) Azure CLI installed: For command-line interactions.

Setting Up Your Azure Machine Learning Workspace

The Azure Machine Learning workspace is the foundational resource for all your ML activities. It provides a centralized place to manage your experiments, models, data assets, compute targets, and more.

Steps to create a workspace:

1. Navigate to the Azure portal.
2. Search for "Machine Learning" and select "Azure Machine Learning".
3. Click "Create".
4. Fill in the required details: Subscription, Resource group, Workspace name, Region, and optionally, select your desired compute options.
5. Review and click "Create".

You can also create a workspace using the Azure CLI:

az ml workspace create -n my-workspace -g my-resource-group -l eastus

Training Your First Model

Once your workspace is set up, you can start training models. Azure Machine Learning supports various training methods, from scripted training to automated ML.

Using Azure ML Studio:

1. Open your Azure Machine Learning workspace in the Azure portal.
2. Click "Launch studio".
3. Navigate to "Experiments" and click "Create new experiment".
4. Upload your data or connect to existing data assets.
5. Define your training script (e.g., a Python script using scikit-learn, TensorFlow, or PyTorch).
6. Configure a compute target (e.g., a Compute Instance or Compute Cluster).
7. Submit your training job.

Deploying Your Model

After training, you'll want to deploy your model to make it accessible for inference.

• Register Model: Register your trained model in the workspace.
• Create Inference Script: Write a script to load your model and perform predictions.
• Create Deployment Configuration: Specify compute resources for your endpoint (e.g., Azure Container Instance for testing or Azure Kubernetes Service for production).
• Deploy: Deploy your model as a web service endpoint.

Example command for deploying to Azure Container Instance (ACI):

az ml service create realtime -f ./aci-deployment.yml --workspace-name my-workspace

Next Steps

Now that you've taken your first steps, here's what you can explore next:

• Data Management Best Practices
• Exploring Compute Options
• Automated Machine Learning (AutoML)
• Model Management and Monitoring
• Integrations with Other Azure Services