I'm following the Python SDK tutorial for training many models: Here a link to the notebook
So everything is working fine, but now I'm interested in the training results. When I check the Azure ML Studio I can see the pipeline steps as followed: training pipeline.
In the outputs directory of mm-models train I do not get the information I need. But if I check the child-jobs I go down in granularity child-jobs.
Here I can see the trained models and their hyper-parameters for example for the partitioned entities which I want to forecast. So for example a specific product in a certain region. models for a partitioned entity
Now what I want now is from the Python SDK (or in general programatically) retrieve the information stored in this child-jobs which I can access via the Azure ML Studio GUI such as cross-validation forecasts or measures for explainability such as these: Azure ML GUI.
I can see that this information is stored in the output folder of the child-jobs: Output + logs
I can't figure out from the Documentation of the Python SDK how to retrieve these results. I also tried to get the results from a child-job via the Rest API following this thread. The output that I got looks like this: error response Additionally I would not know how to get all the IDs from the child jobs from within the python SDK.
I would appreciate any help!