Azure AI Language
An Azure service that provides natural language capabilities including sentiment analysis, entity extraction, and automated question answering.
404 questions
Where to see a more illustrative error message about what went wrong during training/testing in Azure AI Language Custom Classification Model?
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(70.4, 54%, 16%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EMR%3C/text%3E%3C/svg%3E)
Mürüvvet Hasanbaşoğlu
0
Reputation points
We are experimenting with the Custom Classification Model in Azure AI Language Service and when we start experimenting with relatively more training data, we start getting errors during the model training/testing phase. The whole training job gets canceled and the error messages displayed are super generic and therefore doesn't really help to fix the problem...
Would be great to know where and how we can see a more illustrative error description about what went wrong so that we can correct and continue the training job.
{count} votes
-
Amira Bedhiafi 23,096 Reputation points2024-07-25T17:38:57.5566667+00:00 Always start by Azure Portal :
- Logs and Diagnostics: Navigate to the Azure Portal and go to the resource for your AI Language Service. Check the "Logs" section for detailed logs and diagnostic information. This can provide more context on what went wrong during the training/testing phase.
- Activity Log: The Activity Log can show detailed events related to your resources, including failures and their reasons.
If you are using Azure Machine Learning Studio for orchestrating the training, you can find detailed logs and outputs from the training runs in the "Experiments" section. Click on the specific run that failed to view detailed logs and error messages.
Then Azure CLI, if you are comfortable with it :
Use the Azure Command-Line Interface (CLI) to retrieve detailed information about your training jobs. Commands like
az cognitiveservices account listandaz cognitiveservices account showcan help you manage and diagnose issues.Example:
az cognitiveservices account show --name <YourCognitiveServicesAccountName> --resource-group <YourResourceGroupName>can give more insights.Also, if Application Insights is integrated with your AI Language Service, it can provide detailed telemetry and error messages. You can set up custom logging and monitor the detailed performance and failure metrics.
You can also implement custom logging within your training scripts to capture more detailed error messages and diagnostics.
Sign in to comment
Sign in to answer