Unable to analyze all the mappers in the Document Intelligence custom model.

Question

We have built a custom model for BOE entity, the model has been trained successfully. It functions correctly with single-page PDFs, where all mappers are labeled correctly. However, with multiple-page BOEs, not all entries are being labeled.

Answer 1

Hello Niket, Welcome to MS Q&A

To handle multiple-page PDFs in Azure AI Document Intelligence for a custom model, you should note the following:

1. Page Limit: For PDF documents, you can process up to 2,000 pages. However, if you are using a free tier subscription, only the first two pages will be processed.
2. Training Data: When training a custom model, the maximum number of pages for training data is 500 for a custom template extraction model and 50,000 for a custom neural model. Ensure that your training documents are structured consistently to improve accuracy.
3. File Size: The file size for analyzing documents is limited to 500 MB for paid tiers and 4 MB for free tiers.
4. Visual Consistency: Ensure that the documents you use for training present a consistent visual template. This helps in accurately extracting the BOE entity and other labeled data.

By following these guidelines, you can effectively manage multiple-page PDFs in your custom model for BOE entity extraction.

References:

• Document Intelligence custom models
• Document processing models

Please check and let us know if any further questions

Kindly accept answer if it helps

Thanks
Deepanshu