azureml-sdk-for-r
azureml-sdk-for-r copied to clipboard
Support for Pipelines?
I'm looking for pipeline support similar to the Python version. I have 2 customers doing databricks w/ R who would like to use pipelines over top of these steps with datasets etc.
I do not see pipelines on the readme as a supported or planned to support item. Is it supported? If not, can we please ensure it is on the map?
Need ability to support pipeline creations from CLI & .yaml variations for multi step R projects supporting databricks and AML compute.
We now have an RScriptStep for running arbitrary R scripts on AmlCompute. It is available from Python SDK and the CLI.
I will track this as documentation enhancement to add an example of running a pipeline with the CLI
It would be useful to be able to create pipeline steps, pipelines, and pipeline endpoints from the R SDK as well.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed unless the label is removed or if no further activity occurs within 30 days.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed unless the label is removed or if no further activity occurs within 30 days.
Commenting in the hope the rogue bot will allow this to stay open.
Could anyone turn off the boot for this feature request? Creating pipelines from R SDK is really important. Thanks
@lucazav we are fixing the bot, thanks!
I was testing the feature RScriptStep and then, in my case, I need to run an specific docker image, then, I tryed to set custom_docker_image parameter with a path from Docker Hub, but when I run my experiment I got an error "Unable to get image details : Specified base docker image " .. It is importante to mention that if I run a normal EstimatorStep with the same container image, it runs perfectly !!!
I'm facing the same issue @mvenanc has reported. Maybe it's better to open another thread for it. You can find it here.
Is there a plan to introduce Pipelines from the R SDK?
It would be a great feature, really important.
Thank you!