why do i have a trailing comma included in my identification of the field VendorTaxID with the invoice prebuilt model?

Shahar Spencer 60

why do i have a trailing comma included in my identification of the field VendorTaxID with the invoice prebuilt model? for context my invoice is in hebrew.

dupammi 7,390 Reputation points Microsoft Vendor

2024-05-13T08:53:29.2233333+00:00

Hi @Shahar Spencer

Thank you for using the Microsoft Q&A forum.

The trailing comma in the identification of the field VendorTaxID with the invoice prebuilt model could be due to several factors. It can be some syntax error in the code or can be due to less quality in the invoice etc. It's important to ensure that the code is written correctly and follows the syntax rules, and a good quality invoice is provided.

You can refer to the Azure documentation on the invoice prebuilt model to understand the syntax and structure of the code. Here's the link to the documentation:

https://docs.microsoft.com/en-us/azure/cognitive-services/form-recognizer/overview#prebuilt-invoice-model

Additionally, I suggest you check if the invoice is in the correct format and language. The invoice prebuilt model supports invoices in English, French, German, Italian, and Spanish. If the invoice is in a different language, it may not be recognized correctly.

Also please provide more details about the issue, such as the code snippet you were using, the error message you were receiving, and any other relevant information and attach the sample invoice.
Shahar Spencer 60 Reputation points

2024-05-13T09:04:08.1833333+00:00

my invoice is in hebrew. might this be why it happens? how do i make sure that this does not happen in the custom model i will train later?
dupammi 7,390 Reputation points Microsoft Vendor

2024-05-13T10:01:08.0066667+00:00

Hi @Shahar Spencer

If you're looking to train a custom model to recognize Hebrew documents in Azure Form Recognizer, you can use the 2024-02-29-preview version of the API which supports Hebrew and by selecting this API version when creating a new project in the studio, you can train a custom model to recognize Hebrew documents. Once the model is trained, you can use the corresponding analyze API to extract text from Hebrew documents.

You can also utilize the DI SDK or API to interact with the service and train your custom model. When training the model, specify the language parameter as "he". This informs the model that the training data contains Hebrew documents.

Note that the effectiveness of the model's recognition will depend on the quality and variety of the training data. Make sure to provide a diverse and representative set of Hebrew documents to achieve better results.

I hope this helps!!
Shahar Spencer 60 Reputation points

2024-05-19T12:12:51.6233333+00:00

thanks @dupammi

i think the issue is on the level of the the layout model and am not managing to fix the polygon chosen by the layout model before i start tagging in my custom model
Shahar Spencer 60 Reputation points

2024-05-21T12:53:04.4966667+00:00

hi @dupammi

any ideas on how to fix this?

thank you
dupammi 7,390 Reputation points Microsoft Vendor

2024-05-21T13:12:19.9366667+00:00

Hi @Shahar Spencer

It might be more effective to try fixing the trailing comma in your custom model instead of attempting to adjust it in the prebuilt model, as adjusting the prebuilt model can be challenging.

Share via

why do i have a trailing comma included in my identification of the field VendorTaxID with the invoice prebuilt model?