# 1: Randomly change labels to another class ---------------------------------------------------------------
tf_model.update_train_x_y(X_train, Y_train_flipped)
# 2: Include parameters in model.train() to be for ALL_CNN_C for now ---------------------------------------
tf_model.train()
# 2: Include parameters in model.train() to be for ALL_CNN_C for now --------------------------------------
flipped_results[flips_idx, random_seed_idx, 1:] = tf_model.sess.run(
[tf_model.loss_no_reg, tf_model.accuracy_op],
feed_dict=tf_model.all_test_feed_dict)
print('Flipped loss: %.5f. Accuracy: %.3f' %
(flipped_results[flips_idx, random_seed_idx, 1],
flipped_results[flips_idx, random_seed_idx, 2]))
train_losses = tf_model.sess.run(
tf_model.indiv_loss_no_reg, feed_dict=tf_model.all_train_feed_dict)
train_loo_influences = tf_model.get_loo_influences()
# 3 -------------------------------------------------------------------------------------------------------------------#
# # simple test for parallelization (works)
# np.random.RandomState(100)
# arr = np.random.randint(0, 10, size=[200000, 5])
# data = arr.tolist()
# data[:5]
# def howmany_within_range(row, minimum, maximum):
# """Returns how many numbers lie within `maximum` and `minimum` in a given `row`"""
# count = 0
# for n in row:
# if minimum <= n <= maximum:
# count = count + 1
# return count
# update model's training set w/mislabeled labels and re-train
tf_model.update_train_x_y(X_train, Y_train_flipped)
tf_model.train()
# (what does this part do?) this works the same as orig_results
flipped_results[flips_idx, random_seed_idx, 1:] = tf_model.sess.run(
[tf_model.loss_no_reg, tf_model.accuracy_op],
feed_dict=tf_model.all_test_feed_dict)
print('Flipped loss: %.5f. Accuracy: %.3f' %
(flipped_results[flips_idx, random_seed_idx, 1],
flipped_results[flips_idx, random_seed_idx, 2]))
#
train_losses = tf_model.sess.run(
tf_model.indiv_loss_no_reg, feed_dict=tf_model.all_train_feed_dict
) #tf_model.indiv_loss_no_reg from genericNeuralNet.py. Calculates loss for each train data?
train_loo_influences = tf_model.get_loo_influences(
) # tf_model.get_loo_influences from binaryLogisticRegressioWithLBFGS.py this calculates loo influence for each train data?
print(train_loo_influences.shape)
print('\n')
print('\n')
print(train_loo_influences)
print('\n')
print('\n')
#
for checks_idx in range(
num_check_vals
): #num_check_vals = 6 # ACTUAL x-axis of figure 6
np.random.seed(random_seed + 1)
num_checks = int(num_train_examples / 20) * (checks_idx + 1) #
print('### Flips: %s, rs: %s, checks: %s' %
(num_flips, random_seed_idx, num_checks))