Evaluating and Publishing Custom Models Quiz

5 questions Pass: 70% +25 pts (50 with Pro)

Quiz covering Custom Vision Models

What you'll be asked

  1. Which metric in Custom Vision is used to measure the percentage of actual positive predictions that were correct?
  2. After training a Custom Vision model, you notice a low Recall score for a specific tag. What is the most effective way to improve this?
  3. You have published your model to a Prediction Resource. What is required to call the prediction endpoint from an external application?
  4. Why would you adjust the Probability Threshold on the Custom Vision performance tab?
  5. When exporting a Custom Vision model for mobile deployment (e.g., TensorFlow Lite or CoreML), what limitation must you consider regarding model domains?

Answers and explanations come after you submit.