Hyperparameter tuning : Training script command job runs indefinitely

[o-rod]

Module: Perform hyperparameter tuning with Azure Machine Learning

Lab/Demo: 09/ Hyperparameter tuning.ipynb

Task: Tune hyperparameters with a sweep job

Step: Configure and run a command job in Hyperparameter tuning.ipynb

Description of issue
Commnad job runs indefinitely - remains in Preparing status, does not transition to running status, nor complete status

Repro steps:

1. perform all steps given here: https://microsoftlearning.github.io/mslearn-azure-ml/Instructions/09-Hyperparameter-tuning.html
2. Issue will be encountered in the following section of notebook Configure and run a command job, prior to defining sweep and running sweep job.

[JisenRen]

I also get stuck in this notebook. The compute cluster never scales up even if the sweep job is submitted. It even down scales when I start the hyperparameter tuning notebook immediately after the completion of another job.

[Mihai-Ac]

Hello,

I ran through the notebook and didn´t get an error. From several runs, the diabetes-train job ran with an average of 1 min and 30 seconds.
Could you reproduce the issue? if so, is there anything that you did specifically that might have caused this?

Regards,
Mihai

[o-rod]

Hi,

I re-ran the notebook successfully. The train job ran for about 5 min and the sweep job ran for about 10 min.
This issue should be closed.

Thanks,
orod