result[n_backward], correct[n_backward] = compute_accuracy(back_input)
if log_prefix is not None:
- nb_correct = correct[n_forward].sum()
- nb_total = correct[n_forward].size(0)
- back_nb_correct = correct[n_backward].sum()
- back_nb_total = correct[n_backward].size(0)
+ forward_nb_correct = correct[n_forward].sum()
+ forward_nb_total = correct[n_forward].size(0)
+ backward_nb_correct = correct[n_backward].sum()
+ backward_nb_total = correct[n_backward].size(0)
self.logger(
- f"accuracy {log_prefix} {n_epoch} {model.id=} {nb_correct} / {nb_total}"
+ f"forward_accuracy {log_prefix} {n_epoch} {model.id=} {forward_nb_correct} / {forward_nb_total}"
)
self.logger(
- f"back_accuracy {log_prefix} {n_epoch} {model.id=} {back_nb_correct} / {back_nb_total}"
+ f"backward_accuracy {log_prefix} {n_epoch} {model.id=} {backward_nb_correct} / {backward_nb_total}"
)
return result, correct
compute_accuracy(self.train_w_quizzes[:nmax], log_prefix="train")
- result, correct = compute_accuracy(
+ test_result, test_correct = compute_accuracy(
self.test_w_quizzes[:nmax], log_prefix="test"
)
- main_test_accuracy = correct.sum() / correct.size(0)
+ main_test_accuracy = test_correct.sum() / test_correct.size(0)
self.logger(f"main_test_accuracy {n_epoch} {main_test_accuracy}")
##############################
self.save_quizzes(
result_dir,
f"culture_prediction_{n_epoch:04d}_{model.id:02d}",
- quizzes=result[:72],
+ quizzes=test_result[:72],
show_to_be_predicted=True,
- mistakes=correct[:72] * 2 - 1,
+ mistakes=test_correct[:72] * 2 - 1,
)
return main_test_accuracy