Open jaxoftSkyDust opened 2 years ago
Hi @jaxoftSkyDust, thanks for using the template repo. On the issue you raised, it looks like there have been some API changes to the way models are created. When I walked through for a repro, I hit the same error. I also noticed on the Speech Studio site that the model was uploaded but wasn't properly imported (the error message there points to the legacy API usage). I'll need to update the scripts to use the latest Speech CLI version, I'd guess there will be additional tinkering required after. I've filed a work item against me to track this.
I have worked around it by removing baseModel from the specific line jq --arg LOC "$SPEECH_LOCALE" '.values[]|select(.locale==$LOC).self' models.json | tr -d \" > models.txt
The issue however is now that for some reason the model.json created does not contain all models and tries to load more... this results in failing on github but works fine outside github by compiling on bash. Any reasons why this happens?
"@nextLink": "https://***.api.cognitive.microsoft.com/speechtotext/v3.0/models/base?skip=100&top=100"
Following @nextLink https://***.api.cognitive.microsoft.com/speechtotext/v3.0/models/base?skip=100&top=100 ... Following @nextLink https://***.api.cognitive.microsoft.com/speechtotext/v3.0/models/base?skip=100&top=100 ... Done! ERROR: Additional text encountered after finished reading JSON content: ***. Path '', line 1905, position 1.
Hello, When i setup the pipeline and try the first run with: git push origin BASELINE_0 The SpeechTestDataCI will run. I got an error in the step "Get the baseline model" Error: Failed to get the latest baseline model. Possibly re-run all jobs. Error: Process completed with exit code 1.
output is like: SPX - Azure Speech CLI, Version 1.0 Copyright (c) 2020 Microsoft Corporation. All Rights Reserved.
csr.list.kind=models csr.output.json.file=models.json service.config.key= service.config.region= version=1.0.0 x.command=csr.model.list
Listing models ... Listing models ... Done!
{ "values": [] }
Error: Failed to get the latest baseline model. Possibly re-run all jobs. Error: Process completed with exit code 1.
I see also in the previous step (get the benchmark model) tail: cannot open 'my.benchmark.model.url' for reading: No such file or directory NO EXISTING CUSTOM SPEECH MODELS. TEST THE LATEST BASELINE MODEL.
Does anyone know how to solve? Thanks in advance