Why does Azure Auto ML take so much time to train a time series forecasting Model.

Mohammad Khan 1 Reputation point
2022-12-23T10:07:11.517+00:00

I am trying to train a dataset of approx. 30 MB in size to do inventory forecasting and to do that I am using compute cluster as my compute target whose virtual machine size is Standard_DS12_v2 (4 cores, 28 GB RAM, 56 GB disk). Also I am keeping all the configurations to default while preparing a new job. But this training is failing again and again, and it suggests me to increase compute size but I think this compute size is ok for a dataset as small as 30 MB. Please suggest me some solutions to overcome this. Please response as soon as possible.

Thank you.

Azure Machine Learning
Azure Machine Learning
An Azure machine learning service for building and deploying models.
2,965 questions
Azure OpenAI Service
Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
3,237 questions

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. Ramr-msft 17,741 Reputation points
    2022-12-28T01:08:05.093+00:00

    @Mohammad Khan Thanks, Did you try the solutions suggested.
    Increase the experiment timeout when creating a run. The timeout is usually set to prevent experiments from running indefinitely, but you can increase it if you think your experiment will take longer to complete.

    Subsample your dataset to decrease featurization and training time. If your dataset is very large, it can take a long time to process and train the model. One option is to subsample the dataset by selecting a smaller portion of the data to use for training and testing. This can help speed up the experiment, although it may also reduce the accuracy of the model.

    If you are still facing an issue you can send feedback from ml.azure.com as shown below. This will help you to share the details securely and work with an engineer who can provide more insights about the issue that if it can be replicated.
    274369-image.png

    0 comments

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.