Virtual Agent NLU Train and Test the model via Expert Feedback Loop

Ruth Patterson
Tera Expert

After adding correctons via Expert Feedback Loop, Tuning and then clicking to 'Train the model' I get a training error message

 

Error location # (which links to NLU Workbench home page) Type=Intent; Error= Enter utterances that only contain less than 200 characters and less than 25 words. 

 

I've reviewed the default batch test data and removed any utterances >200 characters or 25 words, but still get the error can anyone help?

2 ACCEPTED SOLUTIONS

Brian Bakker
ServiceNow Employee
ServiceNow Employee

@Ruth Patterson 

This is a known defect that has been fixed in the latest version of the NLU Workbench - Advanced Features application.

 

The utterances > 200 chars or 25 words will be in the [ml_labeled_data] table, and should be deleted from this table, as they should not be present.

 

Let me know if you managed to identify these utterances, delete them and you were able to successfully train the NLU Model without this error message.

View solution in original post

Thanks for the quick reply Brian, I'll let you know how we get on asap.

View solution in original post

2 REPLIES 2

Brian Bakker
ServiceNow Employee
ServiceNow Employee

@Ruth Patterson 

This is a known defect that has been fixed in the latest version of the NLU Workbench - Advanced Features application.

 

The utterances > 200 chars or 25 words will be in the [ml_labeled_data] table, and should be deleted from this table, as they should not be present.

 

Let me know if you managed to identify these utterances, delete them and you were able to successfully train the NLU Model without this error message.

Thanks for the quick reply Brian, I'll let you know how we get on asap.