Open michaelmcneilnet opened 1 year ago
@michaelmcneilnet Thanks for your feedback! We will investigate and update as appropriate.
@pritamso Could you please review add comments on this, update as appropriate.
Author of the artilcle @shouryah would help you with this. Thanks!
@michaelmcneilnet Thank you for bringing this to our attention. I've delegated this to content author @Blackmist, who will review it and offer their insightful opinions.
@ssalgadodev can you investigate with the AutoML PM team if this is possible and if we can clarify the upgrade/v2 articles on this?
@michaelmcneilnet No, AutoML doesn't currently support local computes with v2. We'll add a note to the article to call this out.
Are there any plans to support it? It's a shame to have to use an older version of the sdk to get this feature
@michaelmcneilnet we work on the documentation team and don't have much insight into future plans for the product or SDK.
If you'd like to request support for local runs with AutoML in SDK v2, I recommend opening a GitHub issue in the repo that contains the SDK v2 source (https://github.com/Azure/azure-sdk-for-python/tree/azure-ai-ml_1.5.0/sdk/ml/azure-ai-ml). That would bee directly visible by the product group and they may be able to provide more information on plans.
I'm unclear whether AutoML runs can use local compute with v2 SDK. It seems 'normal' runs can https://learn.microsoft.com/en-us/azure/machine-learning/migrate-to-v2-local-runs?view=azureml-api-2 but AutoML experiments can't https://learn.microsoft.com/en-us/azure/machine-learning/how-to-configure-auto-train?view=azureml-api-2#compute-to-run-experiment.
It would be good if that were more explicit in the docs and with an indication of whether this is coming to v2.
Document Details
⚠ Do not edit this section. It is required for learn.microsoft.com ➟ GitHub issue linking.