Model selection metric in Custom Vision
GGG9
21
Reputation points
It is known (https://video2.skills-academy.com/en-us/azure/cognitive-services/custom-vision-service/getting-started-build-a-classifier) that Custom Vision uses k-fold cross validation to evaluate a trained model's performance in the 'Performance' tab, based on precision, recall, and average precision metrics.
Presumably, Custom Vision uses k-fold cross validation to also tune the model's hyper-parameters (~model selection): what metric does it use to perform that task?
Sign in to answer