Question from the TensorFlow test

A condition for stopping the learning phase

This method makes it possible to stop the learning process as soon as all the training data has been processed at least once each.

This method makes it possible to stop the learning process as soon as all the training data are processed at least n times each, with n fixed with an argument.

This method makes it possible to stop the learning process as soon as the performance on the training data has reached a given objective.

This method makes it possible to stop the learning process as soon as the performance on the validation data has reached a given objective.