Loading the data

We will load the .npy files that are produced when generating frame features using a generator. The code ensures that all the input sequences have the same length, padding them with zeros if necessary:

def make_generator(file_list):
    def generator():
        np.random.shuffle(file_list)
        for path in file_list:
            full_path = os.path.join(BASE_PATH, path)
            full_path = full_path.replace('.avi', '.npy')

            label = os.path.basename(os.path.dirname(path))
            features = np.load(full_path)

            padded_sequence = np.zeros((SEQUENCE_LENGTH, 2048))
            padded_sequence[0:len(features)] = np.array(features)

            transformed_label = encoder.transform([label])
            yield padded_sequence, transformed_label[0]
    return generator

In the previous code, we defined a Python closure function—a function that returns another function. This technique allows us to create train_dataset, returning training...