NLU Models throwing error during training

[Srijan]

Suddenly the NLU models are failing to train. We are getting the error message 'Unable to train the model at this time. Please try again after some time.'

The NLU models were getting trained and published successfully till a day before. Any leads on how to solve this or how to debug this issue would be of help.

Thanks.

[nilimadesai]

Hi Srijan,

Most likely it is an issue on the NLU Server and you really may want to try after few minutes or an hour. 

In case there are errors as a result of your model changes, you can check them by enabling Debug Log and reviewing the messages in the log window. Look for a tab called NLU Train. The error should show up after the JSON. 

In case the error does not go away, you may need to open a HI Case so SN can investigate. 

If I have answered your question, please mark my response as correct and/or helpful.

Nilima

[Chris D]

I got that error once, and not to sound like a rehash, but after seeing it twice in a row, I believe I just waited and tried again later (iirc just a few minutes later) and it worked again ¯\_(ツ)_/¯

[Paul Curwen]

I have experienced this also. The solution I found was the following:

1. Try training in Studio NOT under NLU Model Builder > NLU Models. I had issues where Studio would train successfully, but not from the NLU Models form. Weird, but true

2. If that doesn't work then also check the 'm2m_sys_nlu_intent_entity' table and make sure for your model both the intent and entity fields are both populated

3. Next check your vocabulary and synonyms and make sure those are correct as they can cause training loops

[Robert_Card]

Is it possible this is a bigger issue?  We started expriencing the same issue yesterday afternoon and are still getting the message today.  We are in the middle of rolling this out, so its become quite the headache.  Additionally for those who have been using this for a while, is this common?