Implement Custom Language Models AI-102 Practice Questions (Page 11)

Questions 51–55

1. 51foundation · easy

   What is the purpose of recovering a custom language model from a backup?

   • ATo deploy the model to a new region
   • BTo improve the model's accuracy
   • CTo test the model's performance
   • DTo restore the model to a previous state after a failure or data loss
   Reveal Solution Discussion Select an answer first
2. 52foundation · easy

   In a custom question answering knowledge base, what does a question-and-answer pair consist of?

   • AA prompt and a completion
   • BA source URL and a confidence score
   • CAn intent and an entity
   • DA question and its corresponding answer
   Reveal Solution Discussion Select an answer first
4. 53application · medium

   A production custom language model has become corrupted and is returning incorrect predictions. The team has a backup of the model from last week. What should they do to restore service?

   • AContinue using the corrupted model until the next scheduled update.
   • BImport the backup model and deploy it to the prediction endpoint.
   • CDelete the model and create a new one from scratch.
   • DRetrain the model from the original training data.
   Reveal Solution Discussion Select an answer first
5. 54foundation · easy

   After publishing a knowledge base, what does a client application need to call it?

   • AThe training data used to build the knowledge base
   • BThe endpoint URL and an authentication key
   • CAccess to the Azure portal
   • DA backup file of the knowledge base
   Reveal Solution Discussion Select an answer first
6. 55foundation · easy

   Why is it important to test a deployed language model?

   • ATo create a backup of the model
   • BTo verify that the model performs correctly in the production environment
   • CTo add new training data to the model
   • DTo change the model's deployment region
   Reveal Solution Discussion Select an answer first