Question 105032 in Azure Question posted in
The official documentation to get you started can be found here.

AzureML remote data path

I have a MLTable data asset in Azure ML studio that I am trying to access in python and I can’t figure out the structure of the path.

My datastore name is fooddb and my MLTable name is food. The MLTable was created in Microsoft Azure Machine Learning Studio.

training_data_input  = Input(type=AssetTypes.MLTABLE, path="azureml://datastores/fooddb/paths/food")

timeseries_job = automl.forecasting(
    compute="compute",
    training_data=training_data_input,
    experiment_name="salesforecast",
    target_column_name="QTY",
    primary_metric="r2_score",
    n_cross_validations=5,
    enable_model_explainability=True,
    forecasting_settings=forecast_settings
)
Tags:

2

Answers


  1. 0

    Go to azure portal

    Create Azure Machine Learning Studio resource.

    enter image description here

    Create the Resource

    enter image description here

    Goto to Datastores

    enter image description here

    Click on New Datastore

    enter image description here

    Give complete details

    enter image description here

    The Account key will be generated in Storage account, that will be shown later.

    enter image description here

    enter image description here

    enter image description here

    Go to Storage account

    Click on upload.

    enter image description here
    Give the required details

    enter image description here
    Dataset uploaded

    enter image description here

    The key details which we need to give in Datastore creation is from here

    enter image description here

    enter image description here

    enter image description here
    Click Next

    enter image description here

    The Data URL is the files path in datastore that can be used as the external file path.

    Login or Signup to reply.
  2. 0

    To use an AzureML data (Table) asset in a job, you can pass in the asset id…

    # get the data asset
data_asset = ml_client.data.get(name="<asset-name>", version="<version>")

training_data_input  = Input(type=AssetTypes.MLTABLE, path=data_asset.id)

timeseries_job = automl.forecasting(
    compute="compute",
    training_data=training_data_input,
    experiment_name="salesforecast",
    target_column_name="QTY",
    primary_metric="r2_score",
    n_cross_validations=5,
    enable_model_explainability=True,
    forecasting_settings=forecast_settings
)
    Login or Signup to reply.
Please signup or login to give your own answer.

Back To Top
Search