ML Model data output

Question

Hi,

I was running several time series model using Azure automated machine learning, I didn't write any code. After the running was completed, there are some datasets stored in Azure Blob Storage. But I don't know if these files include the prediction results or not because I can't find a right software to open it . I don't need to deploy the model. I just need a plain spreadsheet which contains the result. Why it is so hard? The attachment is the screenshot of the fiels stored in Blob of the model I ran? What do those files mean?
And I just check the running outcome, it shows there is no output dataset. I was so confused! Do I need to change something when I set the model running?
]1

Answer 1

@Marcus Gonzalez I think you might be interested to check your models explainability which tells more about the results of your run and performance of your model. Since you are not planning to deploy your model as a service running explain for all the models will help you choose the best model that can be later used for deployment. The steps to run explain are mentioned in this document.

The files you might be referring to are files that might have been used for processing your data based on the input settings or configuration of your automl run. The files might be different between child runs as these child runs are basically run against an algorithm and different algorithms produce different set of files. Usually, the end user need not refer these files as automl gets the best model for you from all the child runs which is deployed as a web service on ACI or AKS.