Remix/boost.py at master · DependableSystemsLab/Remix · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
import sys
import numpy as np
import pandas as pd
from diversity_utils import *


from scikeras.wrappers import KerasClassifier
from sklearn.ensemble import AdaBoostClassifier


gpus = tf.config.experimental.list_physical_devices('GPU')
for gpu in gpus:
  tf.config.experimental.set_memory_growth(gpu, True)


import argparse

parser = argparse.ArgumentParser(description='Train model with fault params')
parser.add_argument('--dataset', type=str, choices=['mnist', 'cifar10', 'gtsrb', 'pneumonia'], default='mnist')
parser.add_argument('--final_fault', type=str, default="golden")
parser.add_argument('--batch_size', type=int, default=256)
parser.add_argument('--epochs', type=int, default=5)
parser.add_argument('--natural', action='store_true')
parser.add_argument('--ens_size', type=int, default=3)

parser.add_argument('--modelA', type=str, default="ConvNet")

args = parser.parse_args()

dataset = args.dataset
final_fault = args.final_fault
num_epochs = args.epochs
batch_size = args.batch_size

modelA_name = args.modelA

ens_size = args.ens_size


def generate_csv(y_test, predictions_A):
    list_name = ['ground_truth']
    df = pd.DataFrame(columns=list_name)

    df["ground_truth"] = y_test
    df["predicted_A"] = predictions_A

    file_name = "./remix_results/boost-" + str(ens_size)  + "_" + modelA_name + "_" + dataset + "_" +  final_fault + ".csv"
    df.to_csv(file_name)


def main(argv):
    symmetric = not args.natural

    (x_train, y_train), (x_test, y_test) = load_training_data(dataset, final_fault, symmetric)


    input_shape = x_train.shape[1:]
    modelA = get_model_by_name(modelA_name, input_shape)


    modelA_estimator = KerasClassifier(build_fn= modelA, optimizer=tf.keras.optimizers.Adam(), epochs=num_epochs, batch_size=batch_size, verbose=1)

    boosted_ann = AdaBoostClassifier(base_estimator=modelA_estimator, n_estimators=ens_size)
    boosted_ann.fit(x_train, y_train)

    predictions_A = boosted_ann.predict(x_test)
    y_test = y_test.flatten()

    generate_csv(y_test, predictions_A)


if __name__ == "__main__":
    main(sys.argv[1:])