def evaluate(images_filename, labels_filename, names_filename):
images = deep_utils.load_image_data(images_filename)
labels = deep_utils.load_labels(labels_filename, 2)
names = np.load(names_filename)['arr_0']
model = load_model(MODEL_FILENAME)
return model.evaluate(images, labels)
def load_train_datasets(self, fold_id):
ids = deep_utils.load_ids(self.train_id_filenames[fold_id])
data = deep_utils.load_image_data(self.data_filename)
data = np.take(data, ids, axis=0)
labels = deep_utils.load_labels(self.labels_filename, NUM_CLASSES)
labels = np.take(labels, ids, axis=0)
return (data, labels)
def load_pretrain_datasets(self):
data = deep_utils.load_image_data(self.pretrain_data_filename)
labels = deep_utils.load_labels(self.pretrain_labels_filename,
NUM_CLASSES)
return (data, labels)
def compute_number_of_elements(self):
labels = deep_utils.load_labels(self.labels_filename, NUM_CLASSES)
return labels.shape[0]