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. Provide a means to recover code when one of the workspace resources is deleted

    When a core resource for the ML workspace is deleted, the workspace can no longer retrieve any code. Creating a means to be able to resync a new resource and retrieve the code, or implementing a backup would be another solution.

    5 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 →
  2. Show recent workspaces in ML studio in workspace filter pane

    It would be good to show the recent workspaces in workspace filter pane in azure ml studio that would allow users to quickly switch from one workspace to another. There is enough landscape available on this pane too.

    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 →
  3. Make http status codes controllable from your scoring file (e.g. move the AMLResponse from Preview)

    I am using azureml.core.webservice.Webservice and the python SDK.

    Afaik being able to control the HTTP response code of a Webservice is not possible at the moment without AMLResponse which is in Preview and not recommended for production use cases.

    This is because the return of the run() function in the scoring file affects only the body of the response of the deployed webservice but cannot change the header.
    This means that the scoring file can return an Exception but the response will still have the status code 200. I would like to be able to return 4xx statuses like Bad…

    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 →
  4. Gantt view of all experiments execution in Machine Learning Workspace

    To see a Gantt chart of the time consumed by each of the experiments running simultaneously in the ML.

    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 →
  5. Support Named Entity Recognition labeling

    It would be great if can label Named Entities. Ideally we should be able to export it in different formats like BIO-notation,..

    3 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 →
  6. Allow importing data from on-prem databases like Vertica, Oracle etc.

    We are going through a tool selection process for Machine Learning, and we think that Azure Machine learning fits most of our requirements, however one major drawback is that we cannot connect to other on premise databases other than Microsoft SQL server. Allowing multiple data source connectors gives much more flexibility as compared to other tools in the market.

    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 →
  7. Sharing compute instance in members of same developer team

    Actually compute instance can only be used by one member of team, and others developers can't use this.

    A good feature it's all members of a workspace can use compute instance (with the right rights), and make a queue of experiments runnings.

    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 →
  8. Issue with NaN during Export csv filein Export data module

    Hi,
    I am using Export Data module in Machine learning designer to export csv file. Issue is with numeric columns which contains NaN. Result should be empty field/value but exported csv contains NA, which cause troubles in additional file processing.

    1 vote
    Vote
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    1 comment  ·  Flag idea as inappropriate…  ·  Admin →
  9. Support Python 3.6

    Please support Python 3.6

    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 →
  10. Issues with R Package installation on ML Studio , due to old version

    We were trying to run a R program with two libraries ‘tidyverse’ and ‘tabulizer’ in ML studio with Microsoft R open 3.4.4. And looks like the libraries require higher version of R in ML studio.
    Given below is the error message we are getting.

    Can you help us to resolve this issue or support us by providing an alternative option.
    Thank you.

    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 →
  11. 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 →
  12. Love this idea, really!

    Azure Storage Explorer can save you a bit of hassle from I've seen. Great idea really.

    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. can't see error message of "New datastore"

    When you run "New datastore" of the Import data module in Designer of the AML portal, if the datastore creation fails, no error message is displayed.
    Please change this behavior.

    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 →
  14. Be able to write results to SQL tables through Datasets and Datastores

    Currently it doesn't look like it is possible to directly upload the results of a batch Azure ML job to a SQL database that has been registered within the workspace as a datastore. The SQLDataReference class exists, but none of the methods work: https://docs.microsoft.com/en-us/python/api/azureml-core/azureml.data.sqldatareference.sqldatareference?view=azure-ml-py

    It would be great to be able to easily write ML results to SQL databases that the service principal has write access to.

    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. Make Tabular Dataset from_sql_query() method more like SQL alchemy or equivalent

    The SQL syntax for querying Azure ML datastores (eg Azure SQL) is extremely specific and tricky to get right, for example new lines and tabs for readability are not supported, and weirdly nor is the ORDER BY clause. I propose making the query syntax more akin to that supported by the open source sqlalchemy package. Given that many people already use this, it will also make the lift and shift to Azure ML much easier for people, rather than having to rewrite / reformat programmatic query strings.

    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 →
  16. Add segmentation to data labeling tool

    You have bounding box labeling but not anything to help with segmentation.

    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. Segregate a Workspace across multiple projects.

    Allow us to organize 1 workspace in several projects, allowing greater governance and cost sharing between projects.
    In this case, make it possible for us to create project folders where each folder / project has its associated objects, where we can implement access profiles for its objects.
    We would continue sharing processing clusters, but we could segregate registered models, pipelines, datasets, etc. by project folders.

    Today, as it is, we only achieved this segregation by assigning 1 workspace for each project.
    And this strategy of 1 workspace per project will make projects more expensive, because today, a trainning cluster cannot…

    5 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 →
  18. Pass pipeline parameter value into the Execute python script module

    Hi,
    would be useful to be able to pass pipeline parameter value into the Execute python script module. Then we will have general pipeline settings available in custom python scripts.
    Thanks,
    Pavel

    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 →
  19. 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
    Sign in
    (thinking…)
    Sign in with: Microsoft
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    0 comments  ·  Flag idea as inappropriate…  ·  Admin →
  20. Add Script bundle input to Create Python Model module

    Hi,
    would be useful to have Script bundle input as in Execute Python Script for more advanced training.
    Pavel.

    2 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 →
← Previous 1 3 4 5 6 7 8
  • Don't see your idea?

Machine Learning

Categories

Feedback and Knowledge Base