A street-tree failure desk suspects a defect inside a library the framework ships (the syllabus’s example is a defect in PyTorch). Two models that should agree do not, and the official metrics look impossible. Which mitigations should testers pick?
Select an answer to reveal the explanation.
Short Explanation
A defect inside a library the framework ships can make two models that should agree disagree, and the official metrics look impossible. Mitigate that with ML functional performance testing and/or back-to-back testing. It is still not a TensorFlow or PyTorch product exam, not a live canary as the first check, and not a missing-hardware card review.
Full Explanation
A suspected library or framework defect maps to ML functional performance testing and back-to-back testing. A vendor product exam, a canary as the first check, and a card review miss those mitigations.