def test_three_classes():
y = np.array(['a', 'a', 'c', 'b'])
head = head_module.ClassificationHead(name='a')
adapter = head.get_adapter()
adapter.fit_transform(y)
head.config_from_adapter(adapter)
assert head.loss.name == 'categorical_crossentropy'
def test_clf_head_with_2_clases_get_label_encoder():
head = head_module.ClassificationHead(name="a", num_classes=2)
head._encoded = False
head._labels = ["a", "b"]
assert isinstance(
head.get_hyper_preprocessors()[-1].preprocessor, preprocessors.LabelEncoder
)
def test_clf_head_get_multi_label_preprocessor():
head = head_module.ClassificationHead(name="a", multi_label=True)
head._encoded = True
assert isinstance(
head.get_hyper_preprocessors()[0].preprocessor,
preprocessors.MultiLabelEncoder,
)
def test_three_classes():
y = np.array(["a", "a", "c", "b"])
head = head_module.ClassificationHead(name="a")
adapter = head.get_adapter()
adapter.fit_transform(y)
head.config_from_adapter(adapter)
assert head.loss.name == "categorical_crossentropy"
def test_clf_head_get_sigmoid_postprocessor():
head = head_module.ClassificationHead(name="a", multi_label=True)
head._encoded = True
head._encoded_for_sigmoid = True
assert isinstance(
head.get_hyper_preprocessors()[0].preprocessor,
preprocessors.SigmoidPostprocessor,
)
def test_multi_label_loss():
head = head_module.ClassificationHead(name='a', multi_label=True, num_classes=8)
head.output_shape = (8,)
input_node = tf.keras.Input(shape=(5,))
output_node = head.build(kerastuner.HyperParameters(), input_node)
model = tf.keras.Model(input_node, output_node)
assert model.layers[-1].activation.__name__ == 'sigmoid'
assert head.loss.name == 'binary_crossentropy'
def test_clf_head_build_with_zero_dropout_return_tensor():
block = head_module.ClassificationHead(dropout=0, shape=(8,))
outputs = block.build(
keras_tuner.HyperParameters(),
tf.keras.Input(shape=(5,), dtype=tf.float32),
)
assert len(nest.flatten(outputs)) == 1
def test_two_classes():
y = np.array(['a', 'a', 'a', 'b'])
head = head_module.ClassificationHead(name='a')
adapter = head.get_adapter()
adapter.fit_transform(y)
head.config_from_adapter(adapter)
head.output_shape = (1,)
head.build(kerastuner.HyperParameters(), input_module.Input(shape=(32,)).build())
assert head.loss.name == 'binary_crossentropy'
def test_three_classes_infer_categorical_crossentropy():
dataset = np.array(["a", "a", "c", "b"])
head = head_module.ClassificationHead(name="a", shape=(1,))
adapter = head.get_adapter()
dataset = adapter.adapt(dataset, batch_size=32)
analyser = head.get_analyser()
for data in dataset:
analyser.update(data)
analyser.finalize()
head.config_from_analyser(analyser)
head.build(
keras_tuner.HyperParameters(),
input_module.Input(shape=(32,)).build_node(keras_tuner.HyperParameters()),
)
assert head.loss.name == "categorical_crossentropy"
def test_clf_head_hpps_with_uint8_contain_cast_to_int32():
dataset = utils.generate_one_hot_labels(100, 10, "dataset")
dataset = dataset.map(lambda x: tf.cast(x, tf.uint8))
head = head_module.ClassificationHead(shape=(8, ))
analyser = head.get_analyser()
for data in dataset:
analyser.update(data)
analyser.finalize()
head.config_from_analyser(analyser)
assert any([
isinstance(hpp, hyper_preprocessors.DefaultHyperPreprocessor)
and isinstance(hpp.preprocessor, preprocessors.CastToInt32)
for hpp in head.get_hyper_preprocessors()
])
