def main()

in run_miniimagenet.py [0:0]

• 17 lines of code
• 2 McCabe index (conditional complexity)

def main():
    """
    Load data and train a model on it.
    """
    args = argument_parser().parse_args()
    random.seed(args.seed)

    train_set, val_set, test_set = read_dataset(DATA_DIR)
    model = MiniImageNetModel(args.classes, **model_kwargs(args))

    with tf.Session() as sess:
        if not args.pretrained:
            print('Training...')
            train(sess, model, train_set, test_set, args.checkpoint, **train_kwargs(args))
        else:
            print('Restoring from checkpoint...')
            tf.train.Saver().restore(sess, tf.train.latest_checkpoint(args.checkpoint))

        print('Evaluating...')
        eval_kwargs = evaluate_kwargs(args)
        print('Train accuracy: ' + str(evaluate(sess, model, train_set, **eval_kwargs)))
        print('Validation accuracy: ' + str(evaluate(sess, model, val_set, **eval_kwargs)))
        print('Test accuracy: ' + str(evaluate(sess, model, test_set, **eval_kwargs)))