Press "Enter" to skip to content

Category: Machine Learning

Becoming Familiar with MLflow

Published 2022-12-16 by Kevin Feasel

Tomaz Kastrun continues an advent series on Azure ML:

MLFlow is an open-source framework for registering, managing and tracking machine learning models. It is multiplatform, bringing consistent model training and model consumption across different platforms. This means, that training a model locally and uploading it to Azure or training a model on remote compute instances and downloading it, is a great feature for MLflow.

You can use MLflow with Azure CLI, Azure Python SDK or in the studio and it will deliver a consistent experience (note, some functionalities are limited to the language).

Click through for a quick overview of MLflow.

Comments closed

AutoML and Model Registration in AML

Published 2022-12-14 by Kevin Feasel

Tomaz Kastrun continues an advent of Azure Machine Learning. Day 13 covers the topic of Automated ML:

Automated ML is a no-code automated machine learning task. It iterates over many combinations of algorithms and hyperparameters in order to find the best model for your dataset and your prediction variable(s). The final solution is a model, that can be downloaded and later reused. So Automated ML is not just giving you the best model out of a family of algorithms, but lets you use the model, generate the scripts and create the artefacts.

Day 14 concerns model registration:

Important asset is the “Models” in navigation bar. This feature allows you to work with different model types -> custom, MLflow, and Triton. What you do here is, you register a model from different locations (e.g.: local file, AzureML Datastore, AzureML Job, MLflow Job, Model asset in AzureML workspace, and Model asset in AzureML Registry).

Once you open the Models asset, you will see, that you can do many things here. I have already model register from the running the notebook on day4.

Comments closed

Pipelines and Jobs in Azure ML

Published 2022-12-13 by Kevin Feasel

Tomaz Kastrun continues an advent on Azure ML. Day 11 covers pipelines:

A pipeline is set of instructions (or a workflow) for executing particular work of a machine learning task. The idea behind pipelines is that will help the team of data scientists and machine learning engineers standardize workflow and incorporate best practices of preparing data, producing training models, executing the models and deploying them. Pipelines will help improve and build workflow efficiently and in such a way that it can be reusable.

And the idea behind it, is to split a machine learning process into smaller tasks, a multistep workflow, where each step is a separate component than can be developed, upgraded, optimised, configured, automated, and deleted separately. And these steps, connected through interfaces, form a workflow.

Day 12 makes us get a job:

An Azure ML job executes a task against a specified compute target. This is also how the job is created. By configuring a new job, you can also scale out model training, since there are single node and distributed training available.

A simple job command would be to execute a command in a Docker container. And further parameter sweeping can be executed, by specifying it in the job itself. 

Comments closed

C# Text Classification via ML.NET 2

Published 2022-12-12 by Kevin Feasel

Matt Eland tries out ML.NET 2.0:

Recently ML.NET 2.0 was released, giving us a bevy of new features for the open source machine learning library for dotnet applications.

The release improved ML.NET’s text processing capabilities and improved some aspects of the already fantastic automated machine learning capabilities it had. Moreover, the release seemed to reaffirm ML.NET’s determination to be relevant for advanced machine learning tasks, including deep learning and transformer-based architectures.

In this article we’ll explore ML.NET 2.0’s new text classification capabilities and see how you can use C# to analyze sentiment, match utterances to intents, or otherwise classify textual data without having to write a lot of custom code.

Read on to learn more about ML.NET and plenty of turtles.

Comments closed

Working with the AML Python SDK

Published 2022-12-12 by Kevin Feasel

Tomaz Kastrun continues a series on Azure Machine Learning. Day 9 takes us through a piece of the Python SDK:

Python SDK namespace is azureml.core.environment. Environments specify the set of Python packages, environment variables, and software settings around your training and scoring scripts. In addition to Python, you can also configure PySpark, Docker and R for environments.

You can use namespace  Environment (or created object/asset) to make deployment and code reusable for training purposes at given docker images, configurations and compute type.

Day 10 shows us how to work with the Python SDK via VS Code or a local Jupyter notebook:

Let’s continue to explore the power of SDK and the namespaces. And we will look into namespace that will help you connect to Azure ML resources with Python SDK.

Comments closed

AML Environments and SDKs

Published 2022-12-09 by Kevin Feasel

Tomaz Kastrun continues an advent of Azure ML. First up is environments:

We have explored how to create a compute instance and compute target and learned that ML frameworks and scripting packages always come preinstalled.

Choosing the right set of components (CPU, GPU, RAM, Core) and corresponding software (OS, ML Framework, packages) can be time-consuming.

Under Curated environments, you will find predefined environments, with settings for running particular frameworks, like PyTorch or TensorFlow.

Then an overview of the Azure CLI and Python SDK for AML:

What is Azure CLI? It is an Azure Command Line, a great tool for running commands out of CMD. It is a multi-platform and can be run from Azure or from the client’s machine. It is great for scripting and automating repetitive tasks or making the complex task look like lines of code, especially when it comes to infrastructure, managing, provisioning and monitoring. It can also be run from Azure Cloud Shell. It is native to Azure and can be used across all the services and offerings. Usually, the Azure CLI commands start with “az ..”. On top of that, you can also install Azure Machine Learning CLI, as an extension to Azure CLI. The AML CLI will give you additional commands to manage resources for machine learning.

The same functionality (to some extent) in Azure Machine Learning can be achieved with Python SDK. In addition to that, it offers also great ways to create and manage resources you use for training and deployment of models.

And, so that we can catch up a bit to Tomaz, one more post covering the Python SDK:

Looking briefly into Azure CLI and Python SDK, let’s explore the power of SDK and the most important namespaces.

Comments closed

Installing ML Services on SQL Server 2022

Published 2022-12-08 by Kevin Feasel

Tomaz Kastrun notices a change to the SQL Server installer:

Machine Learning Services and language extensions is available under Database Engine Services, and if you want to use any of these languages, check this feature. During the installation process, the R, Python or Java will not be installed (nor asked for permissions), but you will install your own runtime after the installation. This will bring you more convenience with the installation of different R/Python/Java runtimes.

Read on to see how you can install and work with languages like R, Python, and Java in SQL Server 2022.

Comments closed

Data and Compute in Azure ML

Published 2022-12-06 by Kevin Feasel

Tomaz Kastrun continues an advent series on Azure ML. Day 4 covers data sources:

Yesterday, we have learned the general outlook of the Studio and in this blog post, we will be focusing primarily on getting data to the workspace and reading data from other data sources.

Day 5 has you provision some compute:

With a basic understanding of data assets, let’s create compute instances. Under “Manage” in the navigation bar, select “Compute” (denoted as 1), select “Compute instances” (d. 2) and click on “+ New”.

Comments closed

AI: Lying at Card Games via Probabilistic Modeling

Published 2022-12-06 by Kevin Feasel

Matt Eland is on a mission:

I taught an AI to lie at my favorite card game.

In this article, I’ll explore how I did that and what considerations I had to make while designing an artificial intelligence to play a social deduction-based card game. I’ll also discuss where the project is headed and the potential approaches that you might consider building game systems as an AI developer.

The game I chose to model is One Night Ultimate Werewolf by Bezier Games. This is a social deduction game modelled on the popular party games of Werewolf and Mafia. I’ll give you a quick overview of the rules in the next section, for those unfamiliar with it.

This is an interesting overview of the card game, as well as describing the project itself.

Comments closed

Creating an AML Workspace and Trying the Studio

Published 2022-12-05 by Kevin Feasel

Tomaz Kastrun continues an advent of Azure ML. First up, Tomaz creates a workspace:

You will select “New workspace”. For now, we will work on a workspace. But just to mention, the “New registry” will enable you to share assets among different workspaces, support multi-region replication and help you provision all resources to facilitate region replications.

From there, the focus shifts to using Azure Machine Learning Studio:

In this overview page, you can click the button “Launch studio” in the middle of the workspace or you can copy and paste the Studio web URL provided under the “Essentials” to start the Studio.

But before we launch the Studio, let’s explore some additional settings, worth mentioning.

Comments closed