AI-102 Question 405
Select 4You are tasked with automating the processing of invoices for your company using Azure Form Recognizer. The invoices have a custom layout not supported by prebuilt models. You decide to implement a custom document intelligence model. What steps should you take to train and deploy this model successfully?
- A
Label a set of sample documents with the required fields and regions using the Form Recognizer Studio.
- B
Upload the labeled training data to an Azure Blob Storage container connected to the Form Recognizer service.
- C
Enable an Azure Cognitive Search instance to preprocess the documents before training the model.
- D
Train the model using the labeled data in the Form Recognizer Studio or via REST API.
- E
Evaluate the model's performance using test documents and refine the training data if necessary.
Show answer and explanation
Correct answers: A, B, D, E
Explanation
To implement a custom document intelligence model in Azure Form Recognizer, you must provide labeled training data, upload it to Azure Blob Storage, train the model using Form Recognizer Studio or API, and evaluate its performance. Azure Cognitive Search is not relevant to this process, as it is a separate service not required for custom document model training.
- A. Correct.
Correct. Labeling the sample documents with specific fields and regions is essential for training a custom document intelligence model in Azure Form Recognizer.
- B. Correct.
Correct. The training data must be uploaded to Azure Blob Storage and connected to the Form Recognizer service for the model to access it during the training process.
- C. Incorrect.
Incorrect. Azure Cognitive Search is not required for preprocessing documents when creating a custom document intelligence model in Form Recognizer.
- D. Correct.
Correct. Training the model is a critical step, and it can be done using the Form Recognizer Studio or through the REST API.
- E. Correct.
Correct. Evaluating the model's accuracy with test documents and refining the training data ensures the model performs well in production scenarios.