After successfully training the model I get an error when trying to 'Try Model'
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-20-2024 12:22 PM
I have a model that has been successfully trained and published.
When I attempt to 'Try model' and Enter an utterance to test - I get the following error:
Exception caught in the predict call - reason: {"status":"failure","exception":"sn_ml.MLServiceUtil.predict(): java.lang.Exception: ModelInfo artifact not present for key : 5cfe17cc472c02d09fd5f3aa436d43fe_.modelInfo_2"}

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-24-2024 08:02 AM
Indicates the model artifact is missing for this model in table [ml_model_artifact]. Please check that any linked Vocabulary Sources (Table) are synced and then try to train the NLU Model again. If the issue persists, raise a case with Support.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
03-04-2024 06:19 AM
Thanks for your response Brian. I was not able to identify the artifact and all of my vocabulary sources were synced. I opened a case with Support.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
06-19-2024 08:43 AM
Hi Apriled, we are experiencing the same issue. Just wondering what the resolution was from your case with support? Thanks

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
07-01-2024 07:33 AM - edited 07-01-2024 07:36 AM
I found the case and the issue was fixed for April. The root cause to this issue is documented in KB0861056: Clone does not preserve ML Solutions when setting property glide.platform_ml.clone_artifacts to true (Note: You need to be logged into ServiceNow Support to see this KB article).
I suspect it is missing the artifacts for the Vocabulary Source Lists associated to the NLU Model. You can either export the missing attachments on the [ml_model_artifact] records by exporting these records from the source instance to the target instance. Alternatively, you can retrain the NLU Model after removing the "solution_name" value from the [sys_nlu_vocabulary] table, where Type='Static Lookup', as follows -