def generate_train(x_train, y_train, mp_pool):
# this is the augmentation configuration we will use for training
train_datagen = T.ImageDataGenerator(
featurewise_center=False, # set input mean to 0 over the dataset
samplewise_center=False, # set each sample mean to 0
featurewise_std_normalization=False, # divide inputs by std of the dataset
samplewise_std_normalization=False, # divide each input by its std
zca_whitening=False, # apply ZCA whitening
rotation_range=0, # randomly rotate images in the range (degrees, 0 to 180)
width_shift_range=0.2, # randomly shift images horizontally (fraction of total width)
height_shift_range=0.2, # randomly shift images vertically (fraction of total height)
horizontal_flip=True, # randomly flip images
vertical_flip=False, # randomly flip images
zoom_range=[.8, 1],
channel_shift_range=30,
fill_mode='reflect')
train_datagen.config['random_crop_size'] = (299, 299)
train_datagen.set_pipeline([T.random_transform, T.random_crop, T.preprocess_input])
return train_datagen.flow(x_train, y_train, batch_size=batch_size, seed=3, pool=mp_pool)
def generate_test(x_test, y_test, mp_pool):
test_datagen = T.ImageDataGenerator()
test_datagen.config['random_crop_size'] = (299, 299)
test_datagen.set_pipeline([T.random_transform, T.random_crop, T.preprocess_input])
return test_datagen.flow(x_test, y_test, batch_size=batch_size, seed=11, pool=mp_pool)