Why would you manually log your models? In my NN experiments, I automatically write the list of all hyperparameter values and the corresponding performance to a

Why would you manually log your models? In my NN experiments, I automatically write the list of all hyperparameter values and the corresponding performance to a file. In addition, I automatically generate and save graphs showing the results, typically one graph per a nested 'for' loop.