Pediatric-Apnea-Detection/train.py at main · healthylaife/Pediatric-Apnea-Detection · 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
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
import keras
import keras.metrics
import numpy as np
from keras.callbacks import LearningRateScheduler, EarlyStopping
from keras.losses import BinaryCrossentropy
from sklearn.utils import shuffle

from models.models import get_model

THRESHOLD = 1
FOLD = 5


def lr_schedule(epoch, lr):
    if epoch > 50 and (epoch - 1) % 5 == 0:
        lr *= 0.5
    return lr


def train(config, fold=None):
    data = np.load(config["data_path"], allow_pickle=True)
    x, y_apnea, y_hypopnea = data['x'], data['y_apnea'], data['y_hypopnea']
    y = y_apnea + y_hypopnea
    ########################################################################################
    for i in range(FOLD):
        x[i], y[i] = shuffle(x[i], y[i])
        x[i] = np.nan_to_num(x[i], nan=-1)
        if config["regression"]:
            y[i] = np.sqrt(y[i])
            y[i][y[i] != 0] += 2
        else:
            y[i] = np.where(y[i] >= THRESHOLD, 1, 0)

        x[i] = x[i][:, :, config["channels"]]  # CHANNEL SELECTION

    ########################################################################################
    folds = range(FOLD) if fold is None else [fold]
    for fold in folds:
        first = True
        for i in range(5):
            if i != fold:
                if first:
                    x_train = x[i]
                    y_train = y[i]
                    first = False
                else:
                    x_train = np.concatenate((x_train, x[i]))
                    y_train = np.concatenate((y_train, y[i]))

        model = get_model(config)
        if config["regression"]:
            model.compile(optimizer="adam", loss=BinaryCrossentropy())
            early_stopper = EarlyStopping(monitor='val_loss', patience=10, restore_best_weights=True)

        else:
            model.compile(optimizer="adam", loss=BinaryCrossentropy(),
                          metrics=[keras.metrics.Precision(), keras.metrics.Recall()])
            early_stopper = EarlyStopping(monitor='val_loss', patience=10, restore_best_weights=True)
        lr_scheduler = LearningRateScheduler(lr_schedule)
        model.fit(x=x_train, y=y_train, batch_size=512, epochs=config["epochs"], validation_split=0.1,
                  callbacks=[early_stopper, lr_scheduler])
        ################################################################################################################
        model.save(config["model_path"] + str(fold))
        keras.backend.clear_session()


def train_age_seperated(config):
    data = np.load(config["data_path"], allow_pickle=True)
    x, y_apnea, y_hypopnea = data['x'], data['y_apnea'], data['y_hypopnea']
    y = y_apnea + y_hypopnea
    ########################################################################################
    for i in range(10):
        x[i], y[i] = shuffle(x[i], y[i])
        x[i] = np.nan_to_num(x[i], nan=-1)
        if config["regression"]:
            y[i] = np.sqrt(y[i])
            y[i][y[i] != 0] += 2
        else:
            y[i] = np.where(y[i] >= THRESHOLD, 1, 0)

        x[i] = x[i][:, :, config["channels"]]  # CHANNEL SELECTION

    ########################################################################################
    first = True
    for i in range(10):
        if first:
            x_train = x[i]
            y_train = y[i]
            first = False
        else:
            x_train = np.concatenate((x_train, x[i]))
            y_train = np.concatenate((y_train, y[i]))

    model = get_model(config)
    if config["regression"]:
        model.compile(optimizer="adam", loss=BinaryCrossentropy())
        early_stopper = EarlyStopping(monitor='val_loss', patience=10, restore_best_weights=True)

    else:
        model.compile(optimizer="adam", loss=BinaryCrossentropy(),
                      metrics=[keras.metrics.Precision(), keras.metrics.Recall()])
        early_stopper = EarlyStopping(monitor='val_loss', patience=10, restore_best_weights=True)
    lr_scheduler = LearningRateScheduler(lr_schedule)
    model.fit(x=x_train, y=y_train, batch_size=512, epochs=config["epochs"], validation_split=0.1,
              callbacks=[early_stopper, lr_scheduler])
    ################################################################################################################
    model.save(config["model_path"] + str(0))
    keras.backend.clear_session()