Endpoint/Deployment issue resulting in unfound deployment of Custom NER model

Question

I have trained a model using Azure Anguage Studio to perform conversational language understanding tasks by following this guide. When I try to send requests to the model using C# code in Unity, the system fails to find the published project. The error message is below:

RequestFailedException: Job task: 'xxxx' failed with validation errors: ['Cannot find published project: xxxx and deployment: xxxx under provided APIM subscription id.']Status: 400 (Bad Request)ErrorCode: InvalidParameterValue

I have been following this guide to create the text analytics client to access the language resource through an endpoint and API key. However, the system keeps reporting that it can't find the published project, and I suspect this could be due to an incorrect endpoint or a missing step in the model or resource deployment. I retrieved the endpoint by going to the language resource and selecting "Keys and Endpoint" under "Resource Management" in the left panel. For model deployment, I deployed the model in Language Studio but haven't followed any steps to deploy or publish the resource in the Azure portal. May I ask if I have been missing any steps or have gone to the wrong place to copy the endpoint? Any help would be much appreciated!

Answer 1

@Junlong Chen Since you have created a conversational analysis model(CLU) model you will have to use ConversationAnalysisClient to call the endpoint or the model that was trained. Please see this section of the sample from the SDK repo that provides examples to call the model using the client.

If you intend to create a custom text analysis model, you will have to follow the documentation on this page to use it along with the TextAnalyticsClient. I hope this helps!!

If this answers your query, do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.