Create ML : error handling "Data Source Missing"

I have just started using Create ML and discovered it to be unstable and obviously not fully tested.

I created object training data using IBM Cloud Annotation which imported fine and create a model. I then gave Create ML testing data.

However then, in error, I renamed the directory containing the test data. Create ML then entered into, what appears to-be, a loop continuously flashing up a dialogue saying Test Data "Data Source Missing" and there appears to be no way to exist this state other than by closing Create ML. No whenever I open Create ML and navigate to Evaluation > Testing the "Data Source Missing" dialogue starts flashing.

So I now have a Create ML project that appears useless. For this project there is no way to choose test data!

Any ideas?

I should probably raise a bug report.
the resolution, incase anyone is interested, was to duplicate a Model Source, re-add the test data and retrained.

I have the same problem, but in my case I don't be able to end the train and create a model because "data source is missing" appears for the training data before the train ends. I tried few times, I close and reopen the project but I didn't solve. Any solution, please?

In case if data source is out of sync from what you initially provided you have following options

  1. Create a new model source and drag and drop the new data source location

  2. Duplicate current model source and then point to the new data source location

Create ML : error handling "Data Source Missing"
 
 
Q