After we run the classification model, the accuracy from the training data is $90\%,$ and the accuracy from the test data is $70\%.$ Which accuracy should we use to evaluate the performance of this classification model?

Group of answer choices

$90\%$

$70\%$

either one

neither one