Update: Microsoft will be moving away from UserVoice sites on a product-by-product basis throughout the 2021 calendar year. We will leverage 1st party solutions for customer feedback. Learn more here.

Machine Learning

Welcome! The Azure Machine Learning team invites you to share and vote for features to help you build, manage, and deploy custom machine learning models.

Have a technical question, or want to learn more? Please visit our documentation, MSDN forum or StackOverflow.


  1. Azure Policy Support for restricting SKU of AML Compute

    Today, Azure ML Compute does not honor Azure Policy Restrictions on VM SKUs

    I Created an Azure Policy restricting VMs to a specific SKU(VM Size), assigned the Policy to resource group having Azure ML WorkSpace
    Still, I can create ML Compute using AzureML workspace for any size, including the size which is disallowed in the Policy
    Is this a bug or by design? I want to restrict the size of compute being created from AzureML

    5 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  2. Allow for filtering to experiment / error type granularity in Azure Monitoring

    Allow for filtering to experiment granularity in Azure Monitoring.

    Currently metrics only send value 1 if send to Azure Monitoring e.g. if an experiment run succeeds. During experimentation the larger share of experiments will fail, so we need to understand both the error type and the specific experiment at least.

    5 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  3. Remove AKS minimum core restriction ("cluster_purpose")

    If I want to attach an AKS cluster in development environments that don't meet the 4 agents / 12 core requirements, I need to set the "clusterpurpose = AksCompute.ClusterPurpose.DEVTEST" parameter via the Python API. However, this parameter is not supported in ARM templates and also not in the Azure CLI extension, so this complicates our automation strategy.

    The minimum requirements for running an AKS cluster in production are well documented, so it shouldn't be up to the "Machine Learning service" to enforce any restrictions.

    Having this "cluster_purpose" parameter is just an unnecessary complication, and should therefore be removed…

    5 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  4. Need hyper parameter tuning for ML models

    Should allow to tune hyper parameters in order to improve algorithm performance. Implement Grid search, Randomized Search and Hyperopt for parameter tuning.

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  5. Don't silently fail when an ARM template attempts to change an unchangable parameter

    When an ARM template attempts to change a workspace parameter that cannot be changed (such as hbiWorkspace), the template deployment succeeds without any error or notice even though the parameter specified in the template has not taken effect.

    This is counterintuitive at best, and also inconsistent with how other Azure resources work: those do throw an error when you try to change an unchangable parameter using an ARM template.

    Please fix this by always failing the ARM template deployment when the template tries to change an unchangable parameter.

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  6. Need AML SDK for .NET

    AML SDK is only available in Python. I am developing a .Net application (in C#) and need the same features as Python SDK. Is there any plan for porting python SDK to .NET SDK?

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  7. AML Curated Environments

    1) How can I access these for testing?

    2) Is it possible for me to add packages like GDAL or NLTK to a curated environment?

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  8. ML studio Experiments page to allow folders

    In ML Studio, at experiments, we need to allow a user to create folders to organize various projects. Usually one creates different experiments to tackle the problem in different ways. In the end, maybe only one or two experiments are put to use. However, the user would like to retain all prior experiments for future reference on what was tried. It also helps to allow user to add a Readme.txt to go along with each project to document. In our case, a workspace is shared, and the person who created the experiments may leave. In addition, after two years I…

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  9. Azure Machine Learning Studio should be supported for touchscreen use

    Azure Machine Learning Studio requires use of a computer with a mouse for Drag and Drop. One cannot drag using an iPad touch screen for example.
    This may require an app for IOS and a corresponding one for Android devices to enable ML Studio to be supported by devices.

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →

    Hello Azure Customer,

    Thank you for your feedback, your feedback is very important for us and we hear you. This feedback is currently not been prioritized for implementation. As we get more upvotes , we will again access it and may include in our future implementation plans.

    Thank you for understanding.
    Azure CXP Community

  10. Need drop-out parameter for neural network

    The neural network for custom script should put a dropout rate parameter or at least tell us the code for doing dropout in the custom script. Thanks

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  11. Please add function to monitor "Used studio experimentation hour","Used Production Web API's Compute hours about previous month", "Used Prod

    I want to monitor below 3 things.
    Please add function that monitor below 3 things.

    1.Used studio experimentation hour
    2.Used Production Web API's Compute hours about previous month
    3.Used Production Web API's transactions about previous month

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  12. Provide ability to do real Machine Learning in Europe North

    Currently on NCv1 with the ancient K80's that Google do not even give away for free on Google Colab is available in North Europe.
    We have huge amounts of data in NE and thus it would be impossible or at least extremely cost prohibitive to move this data over to WE, where there are V100-NCv3 GPU's.
    Even the NCv3 instances are ridiculously underpowered for SOTA ML work.

    Just to give some perspective on how excruciatingly slow these are to work with:
    My private gaming PC at home running a single 2080ti has 2.5x TensorFlops of a NCv1 running 4 K80…

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →

    Dear Azure Customer,

    Thanks for the feedback. Your feedback is very important to us. Our engineering team will investigate it and consider your feedback seriously. We will get back to you soon.

    Thank you for understanding.
    Azure CXP Community

  13. 4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    3 comments  ·  Flag idea as inappropriate…  ·  Admin →
  14. Script parameters in Azure ML portal

    The RunDetails widget displays the script parameters of the Estimator that were used for that Run. It would be great if they were also displayed on the Azure portal. Furthermore, i would like to be able to search and filter Runs based on the script parameters used.

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  15. Allow for dynamic paths in DataReferences

    Currently, when using the DataReference object to access directories on storage like ADLS Gen 1, we can't dynamically define the path that the DataReference refers to, because it is only evaluated once (at execution time).

    So for instance, if we wanted a DataReference to output data to {CURRENT DATE}/results, such as /DIR/2020/02/10/results, we're not currently able to do this as a datetime.now() would only be called once (when we first evaluate the graph).

    This means we have to do unnecessary extra steps (using the ADLS Gen 1 SDK, or additional ADF steps) to move outputs into the proper directory.

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  16. Azure DevOps ML extension - bug when triggering release pipelines

    I have multiple release pipelines triggered by new model versions. Each pipeline uses different models, but with the same name (they are in different workspaces of the same resource group).

    The issue is that all pipelines seem to be triggered by the same model (the one in the oldest workspace), even though they're using different references, and different service connections. I expect them to be triggered by updates to their respective models, of course.

    I can provide more details if needed.

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  17. Provide predict_proba for multiclass prediction models

    I have successfully trained a multi-class decision forest in ML Studio, but I would like to return all class probabilities as a 2d-array [nscores, nclasses], not just the best class prediction. This is similar to the predict_proba() function from scikit learn models. Can you provide? My customer is asking for this feature.

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  18. Support Python 3.6

    Please support Python 3.6

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  19. restrict access to Storage Account when using ML

    Currently the access to the required storage account cannot be restricted when using ML. For security reasons it is wise to restrict access ("firewalls and virtual networks")

    4 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  20. Allow Properties Shared by Multiple Items to be Edited In a Single Place

    If I select several items in an Azure ML Studio experiement that all share a common property, I should be able to set that property in one place - once - and not have to edit each item separately.

    A huge portion of what data scientist are doing in an experiment is trying out different classification algorithms, training them, scoring them, and evaluating and comparing results. This makes for a lot of repetitive, monotonous work. And it means we have to do exactly the same edits (like setting a dependent variable) across a bunch of different items.

    There's no reason…

    3 votes
    Vote

    We're glad you're here

    Please sign in to leave feedback

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  • Don't see your idea?

Machine Learning

Categories

Feedback and Knowledge Base