A trail-head kiosk transcribes hikers’ spoken reports on a low-power board, but the model was fitted for two weeks on a data-center cluster. How should a tester compare the hardware?
Select an answer to reveal the explanation.
Short Explanation
The data-center cluster fitted the speech model; the low-power trail-head board scores new hiker reports. Training hardware and inference hardware need not be the same class. The kiosk did not train the model, and testers do distinguish those two compute jobs.
Full Explanation
The compute used to create an ML model is often larger than the compute used to score new inputs. Requiring the same cluster, reversing the roles, or saying testers never compare them is incorrect.