Urgent Problem

Question

In azure machine learning how can I upload a binary image classification data set to use in the designer to create a training pipeline ? How should my next follow up components should be ? Convert to image directory ?

Right now I am putting my data set as a zip file that contains two folder a negative folder with the negative images and a positive folder that contains the positive images. However when the pipeline goes to the following block, Convert to image directory, it gives me the following error.

**
We encountered errors while submitting the job**Submit pipelineRun failed. Invalid Graph contains node not support asset while using v2 data assets.

Can anyone help ?

Answer 1

Thanks for the question, you need to upload your binary image classification data set to Azure Machine Learning as a file dataset. A file dataset is a reference to one or more file paths in your datastores or public URLs. File datasets are the only supported dataset type for images.

For the error you are facing A common debugging strategy I have is to delete compute targets and create new ones. Perhaps this was another "transient" error?