trainer_prosup_dtm.py

import numpy as np
import tensorflow as tf
# from tensorflow.keras.models import load_model
from InceptionTime.classifiers.inception import Classifier_INCEPTION
from Data_preparation_Library import system_sleep

gpus = tf.config.list_physical_devices('GPU')
if gpus:
  # Restrict TensorFlow to only use the first GPU
  try:
    tf.config.set_visible_devices(gpus[0], 'GPU')
    logical_gpus = tf.config.list_logical_devices('GPU')
    print(len(gpus), "Physical GPUs,", len(logical_gpus), "Logical GPU")
  except RuntimeError as e:
    # Visible devices must be set before GPUs have been initialized
    print(e)
    
tf.keras.backend.clear_session()

# gpus = tf.config.experimental.list_physical_devices('GPU')
# tf.config.experimental.set_memory_growth(gpus[0], True)

print('Loading Datasets....')
X_train = np.load('prepared_data/X_train1.npy')
X_test = np.load('prepared_data/X_test.npy')
y_train = np.load('prepared_data/y_train1.npy')
y_test = np.load('prepared_data/y_test.npy')

test_X = np.load('prepared_data/test_X.npy')
test_y = np.load('prepared_data/test_y.npy')

def count(y):
    s0 = 0
    s1 = 0
    for arr in y:
        if arr[0]==1:
            s0 = s0+1
        if arr[1]==1:
            s1 = s1+1
    ls = [s0,s1]
    for i in range(2):
        print(i,'-',ls[i],'[',round(ls[i]*100/len(y)),'%]')


for e in [X_train,y_train,X_test,y_test]:
    print(e.shape)

y_true = []
for d in y_test:
    idx = list(d).index(1)
    y_true.append(idx)

class test_loss(tf.keras.callbacks.Callback):

    def __init__(self, X,y):
        super(test_loss, self).__init__()
        self.X_t = X
        self.y_t = y
        
    def on_epoch_end(self, epoch, logs={}):
        logs['test_acc'] = float('-inf')
        test_loss,test_acc = self.model.evaluate(self.X_t,self.y_t,batch_size=1)
        logs['test_acc'] = np.round(test_acc, 5)
        print('test_loss = ',np.round(test_loss, 5),'test_acc = ',np.round(test_acc, 5)) 


clf = Classifier_INCEPTION('', (X_train.shape[1],X_train.shape[2]),nb_classes=2,
                            verbose=True,batch_size=256,nb_epochs=25,nb_filters=32,
                            depth= 5, kernel_size=41)
clf.callbacks = clf.callbacks + [test_loss(test_X,test_y)]

if __name__=='__main__':
    print('Label Distributions :')
    print('Train Data',count(y_train))
    print('Val Data',count(y_test))
    print('Test Data',count(test_y))
    df_metrics = clf.fit(X_train, y_train, X_test, y_test, y_true,plot_test_acc=True)

    system_sleep()