I upload sucessfully data to speech customization (wav audio+ txt transcription) for just one audio in a zip file according to Microsoft docs: https://learn.microsoft.com/en-us/azure/cognitive-services/speech-service/how-to-custom-speech-test-data. When i click to add a test i choose data and it's taking an eternity to process results and never stops processing. My audio is in pt-BR model. Any idea?
I cannot interrupt or deleting tests while it's processing
There is currently an issue in the service, so it doesn't set the status of a test to failed, after we experienced an internal error. A fix for that will be deployed this or next week. After that you should see the test going to a failed state and you can delete it.
Afterwards you should be able to create an accuracy test with the same data and models and it should succeed.
Related
My goal is to collect some custom logs in Azure Monitor from an external VM running on Linux. In that regard, I've installed the log analytics agent according to the MS official documentation, I ran the wizard in order to setup a custom log - that includes a sample file, a row delimiter and a location from where to collect the logs. However, I'm getting a warning message saying:
Two successive configuration applications from OMS Settings failed – please report issue to github.com/Microsoft/PowerShell-DSC-for-Linux/issues (1)
Tried to follow the link proposed that points to Github where I wasn't able to find any solution (nor on any other link) on this and that's why I said to give it a change and ask the community in here.
Though, it is weird that the heartbeat of the machine or manual syslogs messages are being collected except for the custom logs.
Has anyone encountered this and managed to overpass it? Thanks
Apparently, according to the MS answer, the above warning message is normal to be displayed. However, the reason for not collecting the logs was that in the target file that has to be processed by the oms agent, you need to keep appending new entries because this triggers the oms agent which compare and check if the file has new entries than at the last check.
Hope will help someone!
A custom Azure Speech-to-Text test seems to be frozen, cannot be stopped/deleted.
I started a trial with Azure Cognitive Services, custom speech-to-tekst according https://learn.microsoft.com/en-us/azure/cognitive-services/speech-service/how-to-custom-speech-test-data. I succesfully uploaded some .wav files and a UTF-8 BOM encoded text file with one word/utterance per line. Next I created a test project and selected v4.2 unified and V3.0 universal (LM Adapt) as models and started the processing. The test type is inspection and it's status is processing, but that is now for 5 whole days.
I cannot interrupt or delete the test (error message: Delete test failed. Delete 'xxx test1' failed. Error message: The accuracy test did not complete yet.
I seems to be the same issue as this one: Microsoft Cognitive services - Speech customization testing processing seems freezed
Any suggestions on 1) how to stop/delete the test, 2) to get the test finished correctly?
A new release of the Azure Cognitive services speech portal resolved both my issues.
Every time while saving training phrases for intent get following message.
Agent training failed
Internal error occured: 3643ac92-3137-44a0-bb39-ab85ce4c9af0
Agent traning Failed
Internal error occured :1013c13d-625b-4e84-811c-4ec57fd1854b
Any help would be appreciated.
I fixed the issue following the below approach:
Go to Google developer console(https://console.cloud.google.com)
Delete the maps project linked to our agent.
I am still curious about how this worked. Tried different things, but could not come to any conclusion. I guess that's the price I have to pay for using third party NLP service. I integrated maps prebuilt agent, and since then this behavior started appearing. Later removed the map project from the console, and it started working again.
I am curious to know if you have succeeded in resolving yours.
When I try to test my Azure ML model, I get the following error: “Error code: InternalError, Http status code: 500”, so it appears something is failing inside of the machine learning service. How do I get around this error?
I've run into this error before, and unfortunately, the only workaround I found was to create a new ML workspace backed by a storage account that you know is online. Then copy your experiment over to the new workspace, and things should work. It can be a bit cumbersome, but it should get rid of your error message. With the service being relatively new, things sometimes get corrupted as updates are being made, so I recommend checking the box labeled "disable updates" within your experiment. Hope that helps!
I'm brand new to Azure. I'm trying to get a Cloud Service running with 3 web roles.
Last night I created the .cspkg and .cscfg files, exported the certificates and uploaded everything to Azure manager. The manager said my Cloud Service was successful, however for the last 10 hours when I click on "Cloud Services" in the manager it shows my service, but it says "Creating" with wait gif under "SERVICE STATUS".
Is it really still creating? Or did it fail? Is it possible to view more detailed information about the creation process and/or any log files?
Thanks,
Something "bad" has happened. Service Spin up time should be at most a few minutes.
I've seen it take up to 10-15 minutes depending on the hosting center and the number of scaled instances that need to come up but 10 hrs something has definitely gone wrong.
I would delete the service and start again. If you experience the same problem, have a look in your service start up code and make sure that there's no exceptions/infinite loops, other problems in there that might be causing problems.
It may miss any assembly references. If you included any assembly references(packages that are not part of .net) then please ensure that its copy to local attribute is set to true.