How to use Labeled image datasets to perform an image binary classification in Azure ML Designer

Touchard Alexandre 1 Reputation point
2020-06-04T18:07:17.683+00:00

I try to train a model for image binary classification in Azure Machine Learning Designer.

First, I have used the Label Tool, to set a label on each images :

9143-labeltool.jpg

Then, I have exported it as an Azure ML DataSet in order to import it in my ML workflow in the designer, as you can see below :

9087-importdataset.jpg

Then, in order to apply image transformation and to train my model, I have connected my DataSet to a "convert to image directory" module :

9144-workflow.jpg

When I execute the workflow I have the following error :

9145-error.jpg

Also, I could not find any documentations about all of these modules in Azure Machine Learning Designer :

8990-modules-azureml.jpg

So I am not sure about my understanding of how to proceed.

Azure Machine Learning
Azure Machine Learning
An Azure machine learning service for building and deploying models.
2,959 questions
0 comments No comments
{count} votes

1 answer

Sort by: Most helpful
  1. GiftA-MSFT 11,171 Reputation points
    2020-06-05T19:44:03.693+00:00

    Thanks for reaching out. Based on feedback from the product team, the image related modules are being deployed to regions this week and official docs are yet to be published (hopefully by early next week). However, in the meantime, here are some guidance on how to use 'Convert to Image Directory' module.

    Convert to Image Directory (Converts data input to the internal Dataset format used by Microsoft Azure Machine Learning)

    Add the Convert to Image Directory module to your experiment

    Connect the image dataset as input. Supported dataset formats:

    • compressed file in these extensions: '.zip', '.tar', '.gz', '.bz2'
    • folder containing 1 compressed file in above valid extensions
    • folder containing images

    Run the experiment

    Feel free to comment below if you have any further questions or concerns.

    0 comments No 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.