Watson Knowledge Studio "Machine learning model training process failed. Document sets of both type TEST and type BLIND do not exist" error

Resolving The Problem

What';s happening

When you attempt the "Train & Evaluate" or "Evaluate" operations for a custom model in Watson Knowledge Studio, it might fail with the following error:
Machine learning model training process failed. Document sets of both type TEST and type BLIND do not exist

Also, this behavior might be observed in the following situations:

• You do not have any documents assigned to the Test and Blind sets.
  
  
• You have selected a 100/0/0 ratio for Training / Test / Blind as criteria for automatically splitting document sets.

Note: A 100/0/0 ratio for the Training / Test / Blind sets will allow you to execute the "Train" operation. However, it will not allow you to execute the "Evaluate" or combined "Train & Evaluate" operation because the "Evaluate" operation requires Test and Blind document sets.

How to Fix

If you encounter the previously referenced error, complete one of the following actions:

• Allocate a percentage of the documents automatically.
  
  
• Manually assign the documents to the Test and Blind sets.
  
  
• Attempt the "Train" operation only.