def setup_method(self, method):
# super().setup_method(method)
self.train_data = pd.DataFrame(data=np.random.randn(64, 4))
self.val_data = pd.DataFrame(data=np.random.randn(16, 4))
self.test_data = pd.DataFrame(data=np.random.randn(16, 4))
self.past_seq_len = 6
self.future_seq_len_1 = 1
self.future_seq_len_2 = 2
# use roll method in time_sequence
self.feat = TimeSequenceFeatureTransformer()
self.config = {
'batch_size': 32,
'epochs': 1
}
self.model_1 = LSTMSeq2Seq(check_optional_config=False,
future_seq_len=self.future_seq_len_1)
self.model_2 = LSTMSeq2Seq(check_optional_config=False,
future_seq_len=self.future_seq_len_2)
self.fitted = False
self.predict_1 = None
self.predict_2 = None
def test_fit_eval(self):
past_seq_len = 6
future_seq_len = 2
input_dim = 5
output_dim = 4
x_train = np.random.rand(100, past_seq_len, input_dim)
y_train = np.random.rand(100, future_seq_len, output_dim)
x_test = np.random.rand(100, past_seq_len, input_dim)
y_test = np.random.rand(100, future_seq_len, output_dim)
model = LSTMSeq2Seq(check_optional_config=False,
future_seq_len=future_seq_len)
model_config = {
'batch_size': 32,
'epochs': 1,
'latent_dim': 128,
'dropout': 0.2
}
model.fit_eval(x_train, y_train, **model_config)
y_pred = model.predict(x_test)
rmse, smape = model.evaluate(x=x_test, y=y_test, metric=["rmse", "smape"])
assert rmse.shape == smape.shape
assert rmse.shape == (future_seq_len, output_dim)
assert model.past_seq_len == past_seq_len
assert model.future_seq_len == future_seq_len
assert model.feature_num == input_dim
assert model.target_col_num == output_dim
assert y_pred.shape == y_test.shape
def test_predict_with_uncertainty(self, ):
x_train_2, y_train_2 = self.feat._roll_train(
self.train_data,
past_seq_len=self.past_seq_len,
future_seq_len=self.future_seq_len_2)
x_test_2 = self.feat._roll_test(self.test_data,
past_seq_len=self.past_seq_len)
self.model_2.fit_eval((x_train_2, y_train_2), mc=True, **self.config)
prediction, uncertainty = self.model_2.predict_with_uncertainty(
x_test_2, n_iter=2)
assert prediction.shape == (x_test_2.shape[0], self.future_seq_len_2)
assert uncertainty.shape == (x_test_2.shape[0], self.future_seq_len_2)
assert np.any(uncertainty)
new_model_2 = LSTMSeq2Seq(check_optional_config=False)
dirname = tempfile.mkdtemp(prefix="automl_test_feature")
try:
save(dirname, model=self.model_2)
restore(dirname, model=new_model_2, config=self.config)
prediction, uncertainty = new_model_2.predict_with_uncertainty(
x_test_2, n_iter=2)
assert prediction.shape == (x_test_2.shape[0],
self.future_seq_len_2)
assert uncertainty.shape == (x_test_2.shape[0],
self.future_seq_len_2)
assert np.any(uncertainty)
finally:
shutil.rmtree(dirname)
def test_save_restore_2(self):
x_train_2, y_train_2 = self.feat._roll_train(
self.train_data,
past_seq_len=self.past_seq_len,
future_seq_len=self.future_seq_len_2)
x_test_2 = self.feat._roll_test(self.test_data,
past_seq_len=self.past_seq_len)
self.model_2.fit_eval((x_train_2, y_train_2), **self.config)
predict_2_before = self.model_2.predict(x_test_2)
new_model_2 = LSTMSeq2Seq(check_optional_config=False)
dirname = tempfile.mkdtemp(prefix="automl_test_feature")
try:
save(dirname, model=self.model_2)
restore(dirname, model=new_model_2, config=self.config)
predict_2_after = new_model_2.predict(x_test_2)
assert_array_almost_equal(predict_2_before, predict_2_after, decimal=2), \
"Prediction values are not the same after restore: " \
"predict before is {}, and predict after is {}".format(predict_2_before,
predict_2_after)
new_config = {'epochs': 2}
new_model_2.fit_eval((x_train_2, y_train_2), **new_config)
finally:
shutil.rmtree(dirname)