a0351c62548ebddf5c728ec525ad61a4f7827494,tests/hypermodel/test_preprocessor.py,,test_augment,#,111

Before Change

def test_augment():
    raw_images = tf.random.normal([1000, 32, 32, 3], mean=-1, stddev=4)
    augment = preprocessor.ImageAugmentation(seed=5)
    dataset = tf.data.Dataset.from_tensor_slices(raw_images)
    hp = kerastuner.HyperParameters()
    augment.set_hp(hp)
    augment.set_config(augment.get_config())
    for a in dataset:
        augment.transform(a, True)

    def map_func(x):
        return tf.py_function(functools.partial(augment.transform, fit=True),
                              inp=[x],
                              Tout=(tf.float32,))

    new_dataset = dataset.map(map_func)
    for _ in new_dataset:
        pass
    assert isinstance(new_dataset, tf.data.Dataset)


def test_feature_engineering():

After Change

def test_augment():
    dataset = common.generate_data(dtype="dataset")
    new_dataset = run_preprocessor(preprocessor.ImageAugmentation(seed=common.SEED),
                                   dataset,
                                   common.generate_data(dtype="dataset"),
                                   tf.float32)
    assert isinstance(new_dataset, tf.data.Dataset)


def test_feature_engineering():

Instances