def test_call_data_regression(AutoMLFit):
context = an.AutoMLPipeline(an.DataRegression())
context.run_automl()
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=1)
context.run_automl()
assert an.AutoMLFit.is_called
def test_runtime_dataclassifier():
train_file_path = tf.keras.utils.get_file(
"train.csv",
"https://storage.googleapis.com/tf-datasets/titanic/train.csv")
test_file_path = tf.keras.utils.get_file(
"test.csv",
"https://storage.googleapis.com/tf-datasets/titanic/eval.csv")
data_train = pd.read_csv(train_file_path)
data_test = pd.read_csv(test_file_path)
x_train = data_train.drop(columns="survived")
y_train = data_train["survived"]
x_test = data_test.drop(columns="survived")
y_test = data_test["survived"]
context = an.AutoMLPipeline(
an.DataClassification(max_trials=5,
overwrite=True,
loss="mean_squared_error"))
context.run_automl()
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=100)
context.run_automl()
context.train = an.AutoMLEvaluate(x_test, y_test, batch_size=32)
context.run_automl()
context.train = an.AutoMLPredict(x_train, batch_size=32)
context.run_automl()
assert context.return_automl["model"] != None
assert isinstance(context.return_automl["prediction"], np.ndarray)
assert isinstance(context.return_automl["evaluation"], list)
def test_call_timeseries_forecast(AutoMLFit):
context = an.AutoMLPipeline(an.TimeseriesForecaster())
context.run_automl()
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=1)
context.run_automl()
assert an.AutoMLFit.is_called
def test_call_text_classification(AutoMLFit):
context = an.AutoMLPipeline(an.TextClassification())
context.run_automl()
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=1)
context.run_automl()
assert an.AutoMLFit.is_called
def test_call_save(AutoMLFit, AutoMLSave):
context = an.AutoMLPipeline(an.DataClassification())
context.run_automl()
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=1)
context.run_automl()
context.train = an.AutoMLSave("dummy")
context.run_automl()
assert an.AutoMLSave.is_called
def test_call_evaluation(AutoMLFit, AutoMLEvaluate):
context = an.AutoMLPipeline(an.DataClassification())
context.run_automl()
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=1)
context.run_automl()
context.train = an.AutoMLEvaluate(x_test, y_test, batch_size=32)
context.run_automl()
assert an.AutoMLEvaluate.is_called
def test_call_prediction(AutoMLFit, AutoMLPredict):
context = an.AutoMLPipeline(an.DataClassification())
context.run_automl()
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=1)
context.run_automl()
context.train = an.AutoMLPredict(x_train, batch_size=32)
context.run_automl()
assert an.AutoMLPredict.is_called
def test_multi_model():
context = an.AutoMLPipeline(
an.MultiModel(
inputs=[ak.ImageInput(), ak.StructuredDataInput()],
outputs=[
ak.RegressionHead(metrics=["mae"]),
ak.ClassificationHead(loss="categorical_crossentropy",
metrics=["accuracy"]),
],
overwrite=True,
max_trials=2,
))
context.run_automl()
assert context.return_automl["model"] != None
def test_save_load():
data = fetch_california_housing()
x_train, _, y_train, _ = train_test_split(
data.data,
data.target,
test_size=0.33,
random_state=42,
)
context = an.AutoMLPipeline(
an.DataRegression(max_trials=3,
overwrite=True,
loss="mean_squared_error"))
context.run_automl()
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=10)
context.run_automl()
context.train = an.AutoMLSave(model_name="model_autokeras")
context.run_automl()
model = an.AutoMLModels().load_model(model_name="model_autokeras")
assert model != None
def test_runtime_dataregression():
data = fetch_california_housing()
x_train, x_test, y_train, y_test = train_test_split(
data.data,
data.target,
test_size=0.33,
random_state=42,
)
context = an.AutoMLPipeline(
an.DataRegression(max_trials=3,
overwrite=True,
loss="mean_squared_error"))
context.run_automl()
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=10)
context.run_automl()
context.train = an.AutoMLEvaluate(x_test, y_test, batch_size=32)
context.run_automl()
context.train = an.AutoMLPredict(x_train, batch_size=32)
context.run_automl()
assert context.return_automl["model"] != None
assert isinstance(context.return_automl["prediction"], np.ndarray)
assert isinstance(context.return_automl["evaluation"], list)
def test_return_train():
model = an.DataRegression(max_trials=4)
context = an.AutoMLPipeline(model)
assert context.train == model
from sklearn.model_selection import train_test_split
from ai2business.ai_engines import automl_neural_network as an
"""
### Setup the Timeseries Forecaster.
"""
x_train, y_train, x_test, y_test = train_test_split(
dataset.iloc[:, 0:2].values,
dataset.iloc[:, 3].values,
test_size=0.33,
random_state=42,
)
context = an.AutoMLPipeline(an.TimeseriesForecaster())
context.run_automl()
"""
### Fitting the Timeseries Forecaster.
"""
context.train = an.AutoMLFit(x_train, y_train, batch_size=32, epochs=1)
context.run_automl()
"""
### Evaluate the Timeseries Forecaster.
"""
context.train = an.AutoMLEvaluate(x_test, y_test, batch_size=32)
context.run_automl()
