def test_multi_label_deserialize_without_error(): encoder = encoders.MultiLabelEncoder() dataset = tf.data.Dataset.from_tensor_slices([1, 2]).batch(32) encoder = preprocessors.deserialize(preprocessors.serialize(encoder)) assert encoder.transform(dataset) is dataset
def test_multi_label_transform_dataset_doesnt_change(): encoder = encoders.MultiLabelEncoder() dataset = tf.data.Dataset.from_tensor_slices([1, 2]).batch(32) assert encoder.transform(dataset) is dataset
def test_multi_label_postprocess_to_one_hot_labels(): encoder = encoders.MultiLabelEncoder() y = encoder.postprocess(np.random.rand(10, 3)) assert set(y.flatten().tolist()) == set([1, 0])