Azure Machine Learning

Description de Azure Machine Learning

Solution collaborative avec interface glisser-déposer conçue pour créer et déployer des solutions d'analyse prédictive en quelques minutes. Code gratuit.

Informations sur Azure Machine Learning

Microsoft

http://www.microsoft.com

Fondé en 1975

-thumb
-thumb
-thumb

Présentation des tarifs de Azure Machine Learning

Azure Machine Learning est disponible en version gratuite et ne propose pas d'essai gratuit.


Version gratuite

Oui

Essai gratuit

Non

Déploiement et prise en charge de Azure Machine Learning

Ressources d'aide

  • En ligne

Déploiement

  • Cloud, SaaS, web
  • Installation (Windows)

Formation

  • Formation en ligne en direct
  • Documentation

Azure Machine Learning - Fonctionnalités

  • Bibliothèque d'algorithmes de ML
  • Deep learning
  • Modèle de formation
  • Modèles
  • Modélisation prédictive
  • Outils mathématiques et statistiques
  • Traitement automatique des langues (TAL)
  • Visualisation

Logiciels de machine learning : afficher la liste complète

Azure Machine Learning - Alternatives

Plus d'alternatives à Azure Machine Learning

Avis sur Azure Machine Learning

Lire tous les avis

Note globale

4,4/5

Note moyenne

Facilité d'utilisation 3,8
Service client 3,7
Fonctionnalités 4
Rapport qualité-prix 4,2

Évaluer le logiciel

Partagez vos expériences avec d'autres acheteurs de logiciels.

Écrivez votre avis !
Wechuli P.
Software Engineer
Logiciels, 1 001-5 000 employés
Temps d'utilisation du produit: 6 à 12 mois
  • Note globale
    5 /5
  • Facilité d'utilisation
    5 /5
  • Fonctionnalités
    4 /5
  • Support client
    4 /5
  • Rapport qualité-prix
    4 /5
  • Probabilité de recommander le produit
    8/10
  • Source de l'avis 
  • Publié le 07/04/2020

"Machine Learning for Developers"

Commentaires: This is a wonderful service for straightforward machine learning applications such as regression, that don't require deep learning frameworks. It also provides a convenient way of deploying the models thus less time is spent setting up the infrastructure and more time training and tuning the model.

Avantages: - Drag and drop configuration so no need to write actual code. As long as someone has a basic understanding of machine learning concepts, they can configure a working solution.
- Ability to organize projects into separate distinct units, each with its own datasets, experiments and assets
- One-click deployment of the machine learning solution as a service provided one has an Azure account.
- Comprehensive documentation
- You can write custom scripts in R or Python if you need to manipulate the data in some way not provided by the pre-built modules. There is also built in support for Azure Notebooks, so you can attach a notebook anywhere along the workflow, and visualize or manipulate the data in the notebook.

Inconvénients: - The free tier offers a limited number of training hours and functionality so you need to pay a small fee for the service through Azure
- Lacks some specialized deep learning algorithms

  • Source de l'avis 
  • Publié le 07/04/2020
Lewis E.
VP
Temps d'utilisation du produit: 1 à 5 mois
  • Note globale
    4 /5
  • Facilité d'utilisation
    3 /5
  • Fonctionnalités
    4 /5
  • Support client
    2 /5
  • Rapport qualité-prix
    4 /5
  • Probabilité de recommander le produit
    Sans note
  • Source de l'avis 
  • Publié le 07/03/2018

"Some great capabilities here, but the number of machine learning models is limited."

Avantages: The machine learning models that they provide really work well. It is very easy to set up an account and access the possibilities, There are some services that the first usages are free, so you don't chew up money while learning the basics Very fast. Very detailed and clear billing. Allows you to run problems that require very large computer capacity without having to buy very expensive hardware. You pay just for what you use at reasonable rates. Very good overall web descriptions of the capabilities (especially when compared to Amazon, the largest of the providers of this class of software)

Inconvénients: Few models. There are many machine learning models. Azure, Amazon and each of the companies that make offerings only pick a few to provide. Therefore, you can only use for a problem which matches one of their models, or find a way to twist your problem to fit one of the models provided. At this time, this sees to be true of Azure and the others offering cloud-based machine learning. The level of support that Microsoft provides is much less than what they provide to Windows and Office The Sales people know little other than their talking points to advise whether Azure, in it's present state of development, can solve your problem. Moreover, Azure is growing so fast that they cannot hire & train sales people fast enough. It is difficult to estimate the cost of a particular machine learning operation in advance, and whereas Azure is very transparent with the unit prices, it is very hard to estimate how many of the different price element units you will need to run a problem. There is little or no help available now. Only Amazon does better in this area, and Azure is itself not strong here, but is better than all the other non Amazon competitors. While value for the money is terrific & ML is becoming affordable for almost all of us, Azure still seems to be a little more costly than AMS

  • Source de l'avis 
  • Publié le 07/03/2018
Ben W.
Software Engineer
Logiciels, 11-50 employés
Temps d'utilisation du produit: Essai gratuit
  • Note globale
    3 /5
  • Facilité d'utilisation
    2 /5
  • Fonctionnalités
    4 /5
  • Support client
    3 /5
  • Rapport qualité-prix
    3 /5
  • Probabilité de recommander le produit
    5/10
  • Source de l'avis 
  • Publié le 27/09/2019

"Machine Learning for GUI users."

Commentaires: Human pattern recognition, image recognition. Automation

Avantages: When getting into machine learning topics, if you're not a coder, this is likely the option for you. Most traditional ML frameworks require you to code a lot and know a lot of things like convolutions and setting up certain input and outputs. Azure ML Studio makes a lot of that simpler by creating a lot of drag and drop features that can get you going pretty quick.

Inconvénients: Training isn't as straightforward as it may seem. The tutorial videos they have available are a little outdated and there are a few new features not mentioned in them. Also, the free tier is rather limited and is only intended for personal use or some proof-of-concept dev work. I personally feel that Azure ML Studio takes more time to get used to using than learning how to code something like TensorFlow.

  • Source de l'avis 
  • Publié le 27/09/2019
Utilisateur vérifié
Graduate Research Assistant
Enseignement supérieur, 51-200 employés
Temps d'utilisation du produit: 6 à 12 mois
  • Note globale
    5 /5
  • Facilité d'utilisation
    5 /5
  • Fonctionnalités
    4 /5
  • Support client
    Sans note
  • Rapport qualité-prix
    4 /5
  • Probabilité de recommander le produit
    9/10
  • Source de l'avis 
  • Publié le 26/09/2020

"Huge improvements in the last 2 years"

Commentaires: Overall, it has been positive. It's nice having a workspace where I can make VMs and not have to worry about breaking the Azure environment because it's compartmentalized now in Machine Learning.

Avantages: 2 years ago, Azure's Machine Learning (known as Machine Learning Studio) tool option was a drag and drop tool that was slow and limited. Luckily, they got a lot of feedback from data scientists and people using their tools, and they modified their platform a lot so that a data scientist can provision Virtual Machines, connect to databases, and deploy trained machine learning models without needing infrastructure support. It's a great workspace for building models and deploying models now.

Inconvénients: The hardest parts, as usual, is learning specifically how to do something in Azure. If you're coming from AWS, you'll have to relearn how to prototype models and how to deploy them. It's also a little more expensive than standard Azure services, as seems to be the case with machine learning on any platform.

  • Source de l'avis 
  • Publié le 26/09/2020
Utilisateur vérifié
Head of Innovation & Marketing
Logiciels, 51-200 employés
Temps d'utilisation du produit: plus d'un an
  • Note globale
    5 /5
  • Facilité d'utilisation
    3 /5
  • Fonctionnalités
    5 /5
  • Support client
    5 /5
  • Rapport qualité-prix
    5 /5
  • Probabilité de recommander le produit
    8/10
  • Source de l'avis 
  • Publié le 07/03/2018

"With an new machine learning platforms rolling out Azure gave us a head start"

Commentaires: Azure's Machine Learning Studio is a welcome addition to the world of emerging technologies. As a research and development firm we have to experiment of these applications. Azure is a respected product and their foray into machine learning is a welcome one.

Avantages: Microsoft has a solid, reliable and well supported cloud environment that we've been using internally for years. Azure has been a part of what we do for as long as I've been at the company. The price is somewhat okay and the training materials are accessible enough. It's nice to enter the studio to your customized taste and continue to tweak it for maximum efficiency. Database creation/management is where this really shines.

Inconvénients: Pricing isn't the most competitive but we're lucky to be able to ignore that for the most part. Actual support isn't the best but the community does a good job supporting itself.

  • Source de l'avis 
  • Publié le 07/03/2018