How to fix pipeline run failure in Azure AutoML?

Question

How to fix pipeline run failure in Azure AutoML?

KANTEPALLE BHALAJI YADAV 5

I am following the AI-900-AIFundamentals tutorial and encountered an error while creating a classification model. The chart can be found in this link. Can anyone please help me fix this pipeline run failure?

Tutorial link: https://microsoftlearning.github.io/AI-900-AIFundamentals/instructions/02b-create-classification-model.html

romungi-MSFT 49,096 Reputation points Microsoft Employee Moderator

2023-06-28T14:59:39.4433333+00:00

@KANTEPALLE BHALAJI YADAV I am not sure if others on this forum can access your run details, Please check the logs and add more details on what has failed your run.
KANTEPALLE BHALAJI YADAV 5 Reputation points

2023-06-28T16:13:54.5633333+00:00

Apply Transformation std_log.txt
Python std_log (1).txt

Here I am attaching the details, for your reference.

As per the instructions. I have followed the details. Yet! it shows an error saying it doesn't have a Diabetes column.

If I add it in the select column and enter data manually, it doesn't serve the purpose of the model.

Kindly let me know if there is any need for other details.
Thank you for your swift response.
romungi-MSFT 49,096 Reputation points Microsoft Employee Moderator

2023-06-28T17:15:53.24+00:00
@KANTEPALLE BHALAJI YADAV As per the error from Apply Transformation from inference pipeline.

File "/azureml-envs/azureml_8f317849db35f281450cf74333640b98/lib/python3.8/site-packages/azureml/studio/modules/datatransform/scale_and_reduce/normalize_data/nomalize_transformer.py", line 68, in apply ErrorMapping.throw(ColumnNotFoundError(column_name)) > column_name = 'Diabetic'

This might indicate that at step 6 of add transformations included Diabetic column in the training pipeline.

Set the Transformation method to MinMax and the Use 0 for constant columns when checked to True. Edit the columns to transform with Edit columns. Select columns With Rules and copy and paste the following list under include column names: Pregnancies, PlasmaGlucose, DiastolicBloodPressure, TricepsThickness, SerumInsulin, BMI, DiabetesPedigree, Age

Also, at Step 2 & 3 of Add Transformations only patient id needs to be removed and all other columns should be selected.

Find the Select Columns in Dataset module and place it on the canvas, below the diabetes-data dataset. Then connect the output from the bottom of the diabetes-data dataset to the input at the top of the Select Columns in Dataset module. Double click on the Select Columns in Dataset module to access a settings pane on the right. Select Edit column. Then in the Select columns window, select By name and Add all the columns. Then remove PatientID and click Save.

Also ensure that step 5 of add training module of training pipeline added diabetic column as label.

The model we’re training will predict the Diabetic value, so select the Train Model module and modify its settings to set the Label column to Diabetic. The Diabetic label the model will predict is a class (0 or 1), so we need to train the model using a classification algorithm. Specifically, there are two possible classes, so we need a binary classification algorithm.

The error with execute python script is because of the first error which failed Apply transformation and the Score Model module does not have a proper dataset to provide for execute python script. If you are able to fix the problem in the training pipeline and re-run it the inference pipeline should run correctly if the suggested steps are followed.

1 answer

Your answer

romungi-MSFT 49,096 Reputation points Microsoft Employee Moderator

2023-06-28T14:59:39.4433333+00:00

@KANTEPALLE BHALAJI YADAV I am not sure if others on this forum can access your run details, Please check the logs and add more details on what has failed your run.
KANTEPALLE BHALAJI YADAV 5 Reputation points

2023-06-28T16:13:54.5633333+00:00

Apply Transformation std_log.txt
Python std_log (1).txt

Here I am attaching the details, for your reference.

As per the instructions. I have followed the details. Yet! it shows an error saying it doesn't have a Diabetes column.

If I add it in the select column and enter data manually, it doesn't serve the purpose of the model.

Kindly let me know if there is any need for other details.
Thank you for your swift response.

Answer 1

KANTEPALLE BHALAJI YADAV 5

Dear @romungi-MSFT ,
Thank you for your assistance.
The error was in the initial dataset creation where in the initial column header selection.

I chose "All files have the same header" instead of "Only the first file has headers".

A thorough re-work from the initial stage helped me to figure out my mistake over a comparison.

Have a nice day!

Ramkumar 0 Reputation points

2024-05-14T12:54:42.4966667+00:00
@KANTEPALLE BHALAJI YADAV , I tried with "All files have the same header" option and still getting below error. Please Let me know, if you know the root cause

Error Message:

Get ModuleError when invoking ModuleEntry(azureml.studio.modules.python_language_modules.execute_python_script; ExecutePythonScriptModule; run)Traceback (most recent call last):

Share via

How to fix pipeline run failure in Azure AutoML?

1 answer

Your answer