Document Intelligence problem with data collecting using Custom-analyzer (content uderstanding)

Question

Hello,

I have some problem with data collecting. Problem happens when I am collecting data from PDF Invoices which are more than 10 pages long. Mostly I am collecting data from first and last page. The problem is that I am not getting all data from last page. I already corrected it in knowledge and made new model out of it, but still doesn't work well. However when I am reading invoice shorter than 10 pages I dont have any problem. I am calling this service as an api through N8N provider. Could you please give me some tips and tricks how to solve that?

Best regards,
Martin

Answer 1

Hi Palacký Martin

Token count might be exceeding while processing more than 10 pages.  please increase TPM of concerned model in Content understanding.

Create a custom analyzer with Azure Content Understanding in Foundry Tools using REST APIs - Foundr…

If the issue persists even after increasing the TPM.please provide schema, and sample file .

I Hope this helps. Do let me know if you have any further queries.

Thank you!