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.


How can we improve Microsoft Azure Machine Learning?

You've used all your votes and won't be able to post a new idea, but you can still search and comment on existing ideas.

There are two ways to get more votes:

  • When an admin closes an idea you've voted on, you'll get your votes back from that idea.
  • You can remove your votes from an open idea you support.
  • To see ideas you have already voted on, select the "My feedback" filter and select "My open ideas".
(thinking…)

Enter your idea and we'll search to see if someone has already suggested it.

If a similar idea already exists, you can support and comment on it.

If it doesn't exist, you can post your idea so others can support it.

Enter your idea and we'll search to see if someone has already suggested it.

  1. Deploy a web service to AKS without a public endpoint

    It seems like the Machine Learning service currently always uses a public load balancer with a public IP when deploying a web service to AKS.

    We do not want to expose our web services to the public as that is a security risk - even with the Authorization keys. Also, even if we had to expose it, we'd prefer to use our own centralized ingress system to control the traffic coming into the cluster (e.g. via Azure Application Gateway v2).

    We therefore would like to be able to choose one of the following additional methods when connecting to AKS/when deploying…

    16 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  2. Bring back CLI image commands ("az ml image create" etc)

    The earlier preview versions of the CLI extension (e.g. 1.0.10) contained commands for manipulating the Docker images (e.g. "az ml image create container"). This was very useful, because it made it possible to decouple the time consuming task of building the image from the actual deployment to a compute target and allowed verification of the image prior to deploying to production. (In a CI/CD process, image building was part of the "Build" pipeline while webservice deployment was performed in the "Release" pipeline.)

    In the latest CLI extension version (1.0.33.7), the image commands are not present in the CLI, even though…

    22 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  3. Export Azure ML Visual Interface experiments to Python

    To make the visual UI really useful, it would be great to support export of Azure ML Visual Interface experiments to Python.
    The same applies to the Automated ML UI.

    18 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →

    Hi Azure Customer,

    We have the plan for AutoML. But currently it is not in the roadmap for visual interface. We will be watching for additional upvotes for future priority considerations. Thank you for your feedback and understanding.

    Regards,
    Azure CXP Community

  4. Set cloud_roleName in Application Insights

    It's currently hard to distinguish logs in Application insights because requests from different services/deployments all show up with the name "/score". As far as I've seen the only way to identify the service is to use a "startswith" on the "Container Id" customDimension.

    You should therefore set the "cloud_RoleName" field for all logs (requests, exceptions, traces). It should either be set to the service/deployment name, or it should be possible to specify a custom name per deployment (e.g. for allowing us to follow company-wide naming conventions).

    Setting the "cloud_RoleName" also enables other important features in Application Insights, like the Application…

    9 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  5. Open source the Azure ML SDK and CLI extension

    It would be great if the Azure ML SDK and CLI extension were open-sourced and put on GitHub (like several other Azure tools). The user community could help by fixing minor issues and adding improvements and the increased code transparency would help the users learn to use the tools correctly and understand their behavior.

    The Azure ML SDK and CLI extension are implemented in Python, so everyone has access to the source code anyway, but there is no way currently to contribute fixes to the original projects and determining the changes between versions is cumbersome.

    17 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  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

  6. improve new visual interface run execution performance (it's that bad...)

    ml studio was already slow, for the new visual is several times worst, even after creating the default 2 node cluster, things that took one minute on ml studio now go up to 8-10 mins on visual preview?understand the tracking happening below, cluster overhead and all that, but the point is, as it is, it's unusable, some screenshots ...
    https://twitter.com/rquintino/status/1126263499821387776

    8 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  7. Migration from Azure Machine Learning Studio to Visual Interface

    Ideally this should be done automatically for us, but I'd be happy with an export/import flow that lets us migrate manually.

    5 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →

    Hi Azure Customer,

    Our plan is to support convert the Studio experiment to visual interface experiment. Model and web service are currently out of scope. Thank you for your feedback and please keep eyes on it.

    Regards,
    Azure CXP Community

  8. 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
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  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

  9. 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
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  10. Enable GPU support when deploying to Azure Container Instances

    Azure Container Instances can be configured to provide GPU resources to the container (https://docs.microsoft.com/en-us/azure/container-instances/container-instances-gpu). However, there is no way to ask for GPU resources when deploying ML services via the Azure ML SDK (AciWebservice.deploy_configuration) or Azure CLI (az ml service create aci). Even if the image is built with GPU support enabled (ContainerImage.image_configuration(enable_gpu=True,...)), the image does not work properly in ACI because the GPU resources are not present.

    Please extend the SDK and CLI to allow specifying the GPU count and SKU when deploying to ACI.

    10 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →

    Hi Azure Customer,

    Thank you for your feedback. This feature is currently not planned mainly because ACI doesn’t have full GPU support yet. We will follow up with ACI team and expose it when they add full support. At the meantime, we will keep this feedback and vote opening.

    Thank you for your understanding.

    Regards,
    Azure CXP Community

  11. Please, Add Japanease at Azure Machine Learning Home

    I am using azure Machine Learning Studio at Home.
    But it displays only English. I want to use with Japanese.
    So, Please, Add Japanese at Azure Machine Learning Home(https://studio.azureml.net/)

    2 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    triaged  ·  1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  12. Is there a way to attach Data Bricks Cluster which is in a different resource group

    Is there a way to attach Data Bricks Cluster which is in a different resource group

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  13. Limit of experiments in a workspace

    Is there a limit of how many experiments can we create in one workspace? Is there a limit of how many runs can we create in one experiments?

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →

    Hi Azure Customer,

    We do not have limits on pipelines runs or limits on pipelines per workspace at this point in time (not sure if we want to introduce limits as part of the SKU/billing effort).
    We have a limit on the size of a pipeline (30,000 nodes) and the numbers of schedules that will trigger pipelines per month (100,000)

    We will update our document to make it clear. Thank you.

    Regards,
    Azure CXP Community

  14. We have to be able to export experiment in AML Studio so that it can be imported later with another account

    I developed an experiment in AML Studio and deployed it as a web service.

    I would really appreciate it if I can I export it so that it can be imported later with another Azure Account and it can be redeployed again (in case that i don't have my account anymore or i delete my current experiment)

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  15. Underlying storage account for Machine Learning Service Workspace should be HTTPS only

    Currently the MLSW triggers Azure security toolkit & security center alarms because the default configuration of underlying storage account allows http traffic.

    3 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    triaged  ·  1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  16. Share notebook VM

    Is it possible for 2 users of an AML Workspace to access the same notebook VM? We are both collaborators in the workspace, but my colleague cannot access the notebook VM I created.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  17. Add sklearn Pipeline for feature engineering chaining

    When doing repeated cross validation like k-fold, some feature transformer like normaliser/maxminscaler/feature_selector, etc should be fitted only on the training data (and then tranform the training and validation set). Otherwise some data leakage occurs.

    Therefore, would it be possible to add the sklearn pipeline object as an argument ?
    After splitting the dataset into training and validation set, automl would fit on the training set and apply the pipeline on the validation set, whatever the cv strategy

    One could then add the whole dataset and use the available cross validations strategies without

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  18. Please add function that notify some email when Experiment's Runs is failed

    I hope a function that notify some email when Experiment's Runs is failed.
    I can know whether there is error or not if I open their Experiment manually now.
    But when I run experiment automatically, I want to notify to some email if experiment's Run is failed.

    4 votes
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  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

  19. Add PowerShell to Azure Notebooks

    Please add and improve the kernel for Windows PowerShell/Powershell Core/6/7 to Azure Notebooks.
    PowerShell should be treated as first citizen language and there are a lot of usecases for teaching with Azure Notebooks.
    Alternatively provide an option to load custom kernels.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  20. Continue other paths on error

    When an error occurs on one path it would be nice if the experiment would not stop immediately but continue on other paths and finish as many modules as possible before stopping.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
← Previous 1 3
  • Don't see your idea?

Machine Learning

Feedback and Knowledge Base