def test_cast_to_int32_return_int32():
dataset = utils.generate_one_hot_labels(100, 10, "dataset")
dataset = dataset.map(lambda x: tf.cast(x, tf.uint8))
dataset = common.CastToInt32().transform(dataset)
for data in dataset:
assert data.dtype == tf.int32
break
def test_io_api(tmp_path):
num_instances = 20
image_x = utils.generate_data(num_instances=num_instances, shape=(28, 28))
text_x = utils.generate_text_data(num_instances=num_instances)
image_x = image_x[:num_instances]
structured_data_x = (pd.read_csv(utils.TRAIN_CSV_PATH).to_numpy().astype(
np.unicode)[:num_instances])
classification_y = utils.generate_one_hot_labels(
num_instances=num_instances, num_classes=3)
regression_y = utils.generate_data(num_instances=num_instances,
shape=(1, ))
# Build model and train.
automodel = ak.AutoModel(
inputs=[ak.ImageInput(),
ak.TextInput(),
ak.StructuredDataInput()],
outputs=[
ak.RegressionHead(metrics=["mae"]),
ak.ClassificationHead(loss="categorical_crossentropy",
metrics=["accuracy"]),
],
directory=tmp_path,
max_trials=2,
tuner=ak.RandomSearch,
seed=utils.SEED,
)
automodel.fit(
[image_x, text_x, structured_data_x],
[regression_y, classification_y],
epochs=1,
validation_split=0.2,
batch_size=4,
)
def test_io_api(tmp_path):
num_instances = 100
(image_x, train_y), (test_x, test_y) = mnist.load_data()
(text_x, train_y), (test_x,
test_y) = utils.imdb_raw(num_instances=num_instances)
image_x = image_x[:num_instances]
text_x = text_x[:num_instances]
structured_data_x = utils.generate_structured_data(
num_instances=num_instances)
classification_y = utils.generate_one_hot_labels(
num_instances=num_instances, num_classes=3)
regression_y = utils.generate_data(num_instances=num_instances,
shape=(1, ))
# Build model and train.
automodel = ak.AutoModel(
inputs=[ak.ImageInput(),
ak.TextInput(),
ak.StructuredDataInput()],
outputs=[
ak.RegressionHead(metrics=['mae']),
ak.ClassificationHead(loss='categorical_crossentropy',
metrics=['accuracy'])
],
directory=tmp_path,
max_trials=2,
seed=utils.SEED)
automodel.fit([image_x, text_x, structured_data_x],
[regression_y, classification_y],
epochs=1,
validation_split=0.2)
def test_structured_clf_fit_call_auto_model_fit(fit, tmp_path):
auto_model = ak.StructuredDataClassifier(directory=tmp_path, seed=utils.SEED)
auto_model.fit(
x=utils.generate_structured_data(num_instances=100),
y=utils.generate_one_hot_labels(num_instances=100, num_classes=3))
assert fit.is_called
def test_clf_head_transform_df_to_dataset():
adapter = output_adapter.ClassificationHeadAdapter(name="a")
y = adapter.fit_transform(
pd.DataFrame(utils.generate_one_hot_labels(dtype="np", num_classes=10))
)
assert isinstance(y, tf.data.Dataset)
def test_img_clf_fit_call_auto_model_fit(fit, tmp_path):
auto_model = ak.ImageClassifier(directory=tmp_path, seed=utils.SEED)
auto_model.fit(x=utils.generate_data(num_instances=100, shape=(32, 32, 3)),
y=utils.generate_one_hot_labels(num_instances=100,
num_classes=10))
assert fit.is_called
def test_infer_ten_classes():
adapter = output_adapter.ClassificationHeadAdapter(name="a")
adapter.fit_transform(
utils.generate_one_hot_labels(dtype="dataset", num_classes=10)
)
assert adapter.num_classes == 10
def test_image_classifier(tmp_path):
train_x = utils.generate_data(num_instances=100, shape=(32, 32, 3))
train_y = utils.generate_one_hot_labels(num_instances=100, num_classes=10)
clf = ak.ImageClassifier(directory=tmp_path, max_trials=2, seed=utils.SEED)
clf.fit(train_x, train_y, epochs=1, validation_split=0.2)
keras_model = clf.export_model()
clf.evaluate(train_x, train_y)
assert clf.predict(train_x).shape == (len(train_x), 10)
assert isinstance(keras_model, tf.keras.Model)
def test_clf_head_one_hot_shape_error():
adapter = output_adapter.ClassificationHeadAdapter(name="a", num_classes=9)
with pytest.raises(ValueError) as info:
adapter.fit_transform(
utils.generate_one_hot_labels(dtype="np", num_classes=10)
)
assert "Expect one hot encoded labels to have shape" in str(info.value)
def test_infer_ten_classes():
analyser = output_analysers.ClassificationAnalyser(name="a")
dataset = utils.generate_one_hot_labels(dtype="dataset", num_classes=10)
for data in dataset:
analyser.update(data)
analyser.finalize()
assert analyser.num_classes == 10
def test_structured_clf_fit_call_auto_model_fit(fit, tmp_path):
auto_model = ak.StructuredDataClassifier(directory=tmp_path,
seed=utils.SEED)
auto_model.fit(
x=pd.read_csv(utils.TRAIN_CSV_PATH).to_numpy().astype(
np.unicode)[:100],
y=utils.generate_one_hot_labels(num_instances=100, num_classes=3),
)
assert fit.is_called
def test_clf_head_one_hot_shape_error():
analyser = output_analysers.ClassificationAnalyser(name="a", num_classes=9)
dataset = tf.data.Dataset.from_tensor_slices(
utils.generate_one_hot_labels(dtype="np", num_classes=10)).batch(32)
with pytest.raises(ValueError) as info:
for data in dataset:
analyser.update(data)
analyser.finalize()
assert "Expect the target data for a to have shape" in str(info.value)
def test_functional_api(tmp_path):
# Prepare the data.
num_instances = 80
(image_x, train_y), (test_x, test_y) = mnist.load_data()
(text_x, train_y), (test_x, test_y) = utils.imdb_raw()
(structured_data_x, train_y), (test_x, test_y) = utils.dataframe_numpy()
image_x = image_x[:num_instances]
text_x = text_x[:num_instances]
structured_data_x = structured_data_x[:num_instances]
classification_y = utils.generate_one_hot_labels(
num_instances=num_instances, num_classes=3)
regression_y = utils.generate_data(num_instances=num_instances,
shape=(1, ))
# Build model and train.
image_input = ak.ImageInput()
output = ak.Normalization()(image_input)
output = ak.ImageAugmentation()(output)
outputs1 = ak.ResNetBlock(version='next')(output)
outputs2 = ak.XceptionBlock()(output)
image_output = ak.Merge()((outputs1, outputs2))
structured_data_input = ak.StructuredDataInput()
structured_data_output = ak.CategoricalToNumerical()(structured_data_input)
structured_data_output = ak.DenseBlock()(structured_data_output)
text_input = ak.TextInput()
outputs1 = ak.TextToIntSequence()(text_input)
outputs1 = ak.Embedding()(outputs1)
outputs1 = ak.ConvBlock(separable=True)(outputs1)
outputs1 = ak.SpatialReduction()(outputs1)
outputs2 = ak.TextToNgramVector()(text_input)
outputs2 = ak.DenseBlock()(outputs2)
text_output = ak.Merge()((outputs1, outputs2))
merged_outputs = ak.Merge()(
(structured_data_output, image_output, text_output))
regression_outputs = ak.RegressionHead()(merged_outputs)
classification_outputs = ak.ClassificationHead()(merged_outputs)
automodel = ak.AutoModel(
inputs=[image_input, text_input, structured_data_input],
directory=tmp_path,
outputs=[regression_outputs, classification_outputs],
max_trials=2,
tuner=ak.Hyperband,
seed=utils.SEED)
automodel.fit((image_x, text_x, structured_data_x),
(regression_y, classification_y),
validation_split=0.2,
epochs=1)
def test_structured_data_classifier(tmp_path):
num_data = 500
num_train = 400
data = pd.read_csv(utils.TRAIN_CSV_PATH).to_numpy().astype(np.unicode)[:num_data]
x_train, x_test = data[:num_train], data[num_train:]
y = utils.generate_one_hot_labels(num_instances=num_data, num_classes=3)
y_train, y_test = y[:num_train], y[num_train:]
clf = ak.StructuredDataClassifier(
directory=tmp_path, max_trials=1, seed=utils.SEED
)
clf.fit(x_train, y_train, epochs=2, validation_data=(x_train, y_train))
clf.export_model()
assert clf.predict(x_test).shape == (len(y_test), 3)
def test_structured_data_from_numpy_classifier(tmp_path):
num_data = 500
num_train = 400
data = utils.generate_structured_data(num_data)
x_train, x_test = data[:num_train], data[num_train:]
y = utils.generate_one_hot_labels(num_instances=num_data, num_classes=3)
y_train, y_test = y[:num_train], y[num_train:]
clf = ak.StructuredDataClassifier(directory=tmp_path,
max_trials=1,
seed=utils.SEED)
clf.fit(x_train, y_train, epochs=2, validation_data=(x_train, y_train))
clf.export_model()
assert clf.predict(x_test).shape == (len(y_test), 3)
def test_structured_classifier(init, fit, tmp_dir):
num_data = 500
train_x = utils.generate_structured_data(num_data)
train_y = utils.generate_one_hot_labels(num_instances=num_data, num_classes=3)
clf = structured_data.StructuredDataClassifier(
column_names=utils.COLUMN_NAMES_FROM_NUMPY,
directory=tmp_dir,
max_trials=1,
seed=utils.SEED)
clf.fit(train_x, train_y, epochs=2, validation_data=(train_x, train_y))
assert init.called
assert fit.called
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()
])
def test_text_and_structured_data(tmp_path):
# Prepare the data.
num_instances = 80
(x_text, y_train), (x_test, y_test) = utils.imdb_raw()
x_structured_data = pd.read_csv(utils.TRAIN_CSV_PATH)
x_text = x_text[:num_instances]
x_structured_data = x_structured_data[:num_instances]
y_classification = utils.generate_one_hot_labels(
num_instances=num_instances, num_classes=3)
y_regression = utils.generate_data(num_instances=num_instances,
shape=(1, ))
# Build model and train.
structured_data_input = ak.StructuredDataInput()
structured_data_output = ak.CategoricalToNumerical()(structured_data_input)
structured_data_output = ak.DenseBlock()(structured_data_output)
text_input = ak.TextInput()
outputs1 = ak.TextToIntSequence()(text_input)
outputs1 = ak.Embedding()(outputs1)
outputs1 = ak.ConvBlock(separable=True)(outputs1)
outputs1 = ak.SpatialReduction()(outputs1)
outputs2 = ak.TextToNgramVector()(text_input)
outputs2 = ak.DenseBlock()(outputs2)
text_output = ak.Merge()((outputs1, outputs2))
merged_outputs = ak.Merge()((structured_data_output, text_output))
regression_outputs = ak.RegressionHead()(merged_outputs)
classification_outputs = ak.ClassificationHead()(merged_outputs)
automodel = ak.AutoModel(
inputs=[text_input, structured_data_input],
directory=tmp_path,
outputs=[regression_outputs, classification_outputs],
max_trials=2,
tuner=ak.Hyperband,
seed=utils.SEED,
)
automodel.fit(
(x_text, x_structured_data),
(y_regression, y_classification),
validation_split=0.2,
epochs=1,
)
def test_infer_num_classes():
y = utils.generate_one_hot_labels(dtype='dataset')
adapter = output_adapter.ClassificationHeadAdapter(name='a')
y = adapter.fit(y)
assert adapter.num_classes == 10
def test_tf_dataset():
y = utils.generate_one_hot_labels(dtype='dataset')
head = output_adapter.ClassificationHeadAdapter(name='a')
y = head.fit_transform(y)
assert isinstance(head.transform(y), tf.data.Dataset)
