question

theerthaPadman-2661 avatar image
0 Votes"
theerthaPadman-2661 asked BarnettIan-7170 edited

Form recognizer analysis error V 2.0

I am getting the following error while running the python script to analyse a trained model with labels.

"The Form Recognizer API version used to create the model and the current API version used for analyze are incompatible.

azure-form-recognizer
· 2
5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

@PramodValavala-MSFT : please help me with this question
I am getting the following error while running the python script to analyse a trained model with labels.

"The Form Recognizer API version used to create the model and the current API version used for analyze are incompatible.

0 Votes 0 ·

@theerthaPadman-2661 Did you train the model using the preview v2.1 version?

0 Votes 0 ·
romungi-MSFT avatar image
0 Votes"
romungi-MSFT answered

@theerthaPadman-2661 If the model is trained with a different version than the one used for analyze API this message would be seen. You can use the same version of analyze API to ensure the response is successful.

5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.

BarnettIan-7170 avatar image
2 Votes"
BarnettIan-7170 answered BarnettIan-7170 edited

@theerthaPadman-2661 I also ran into this issue

I am using the current NPM SDK (v3.0.0) to access FormRecognizer, which targets the service API v2.0

However when reading the docs here: https://docs.microsoft.com/en-us/azure/cognitive-services/form-recognizer/quickstarts/label-tool?tabs=v2-0 I found that there are two links with the same text (FOTT Website) but each has a different link value - this is where things went wrong for me:
- https://fott-preview.azurewebsites.net/
- https://fott.azurewebsites.net/

I was using the first one which is v2.1 (preview) so my API calls were failing expecting models trained using v2.0
I switched to the non-preview FOTT Tool which is v2.0 and my API calls began working just fine (although the recognition accuracy drops off notably in v2.0)

All the best!


5 |1600 characters needed characters left characters exceeded

Up to 10 attachments (including images) can be used with a maximum of 3.0 MiB each and 30.0 MiB total.