Microsoft Learn Explore regression with Azure Machine Learning Designer - Inference Pipeline

brandon adleman 0

I am following the Microsoft Learning lab https://microsoftlearning.github.io/AI-900-AIFundamentals/instructions/02a-create-regression-model.html and am getting an error User's image

User's image

I think the lab is several versions as Azure old and something is missing

romungi-MSFT 42,206 Reputation points Microsoft Employee

2023-05-12T05:56:25.9933333+00:00

@brandon adleman Do you see any additional details in the logs of the job? I think the failure could be a missing column value for apply transformation. The logs should provide additional context on what could be missing. Thanks!

brandon adleman 0

@romungi-MSFT sorry for the late reply! This is the log error:

Execution failed. User process '/azureml-envs/azureml_8f317849db35f281450cf74333640b98/bin/python' exited with status code 1. Please check log file 'user_logs/std_log.txt' for error details. Error:   File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/core/logger.py", line 214, in wrapper
    ret = func(*args, **kwargs)
  File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/modulehost/module_host_executor.py", line 52, in execute_with_cli
    do_execute_with_env(parser, FolderRuntimeEnv())
  File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/modulehost/module_host_executor.py", line 63, in do_execute_with_env
    ModuleReflector(parser.module_entry, env).exec(
  File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/modulehost/module_reflector.py", line 397, in exec
    self._handle_exception(bex)
  File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/modulehost/module_reflector.py", line 471, in _handle_exception
    raise exception
  File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/modulehost/module_reflector.py", line 379, in exec
    output_tuple = self._entry.func(**reflected_input_ports, **reflected_parameters)
  File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/modulehost/module_reflector.py", line 76, in wrapper
    ret = func(*args, **validated_args)
  File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/modules/ml/score/apply_transformation/apply_transformation.py", line 51, in run
    return transform.apply(data),
  File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/modules/datatransform/clean_missing_data/clean_missing_transform.py", line 78, in apply
    raise ColumnNotFoundError(
azureml.studio.common.error.ColumnNotFoundError: Column with name or index "price" does not exist in "Dataset", but exists in "Transformation".

romungi-MSFT 42,206 Reputation points Microsoft Employee

2023-05-22T07:25:52.9366667+00:00

@brandon adleman I think you might have added this column in the normalized data module in your training pipeline. The 4th step of Add and explore a dataset section advises to remove the price column and the 5th step of Add data transformations does not have this column. I think you might have missed one of these steps and this is causing the transformation module in the inference pipeline to still use it. Please cross check and try to re-run the training pipeline and then create an inference pipeline.