Trained on your labelled documents, sometimes only five of them, to extract the fields you specify; Document Intelligence offers it as either custom neural or custom template.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Custom extraction model in context, with comparison tables and the common traps.
Terms in this definition
- ELT
Extract, load, transform: raw data lands in the target system first and is transformed there. See ETL for the opposite order.
- Azure Document Intelligence in Foundry Tools
Foundry Tool that pulls text and fields out of forms and documents using read, layout, prebuilt models (such as invoice and receipt) and custom models.
Related terms
- Custom neural model
Custom extraction model in Document Intelligence based on deep learning that copes with differing layouts of the same document type, structured or semi-structured; Microsoft recommends starting with it rather than custom template.
- Custom template model
Custom extraction model in Document Intelligence for documents with one fixed layout, trained in 1-5 minutes; any variation in layout calls for a separate model.