Custom Document Models Quiz

What you'll be asked

1. What is the primary purpose of using a Custom Document Model in Foundry compared to using a pre-trained generic OCR model?
2. When building a Custom Document Model, what is the significance of the 'ground truth' dataset?
3. You notice that your Custom Document Model is failing to extract fields from documents that have a slightly different orientation or rotation. What is the best practice to improve model robustness?
4. In Foundry's Document Intelligence workflow, what is the primary role of the 'Human-in-the-loop' (HITL) review task?
5. You are training a model to extract line items from complex, multi-page invoices. After training, the model shows high precision but low recall for specific line items. What architectural adjustment should you prioritize?

Answers and explanations come after you submit.