def test_extra_trees_classifier(self):
model = ExtraTreesClassifier(n_estimators=3)
dump_one_class_classification(
model,
allow_failure="StrictVersion(onnx.__version__)"
" < StrictVersion('1.2') or "
"StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')",
)
dump_binary_classification(
model,
allow_failure=(
"StrictVersion(onnx.__version__) < StrictVersion('1.2') or "
"StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')"
),
)
dump_multiple_classification(
model,
# Operator cast-1 is not implemented in onnxruntime
allow_failure=(
"StrictVersion(onnx.__version__) < StrictVersion('1.2') or "
"StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')"
),
)
def test_decision_tree_classifier(self):
model = DecisionTreeClassifier()
dump_one_class_classification(model)
dump_binary_classification(model)
dump_multiple_classification(model)
dump_multiple_classification(model, label_uint8=True)
dump_multiple_classification(model, label_string=True)
def test_gradient_boosting_classifier1Deviance(self):
model = GradientBoostingClassifier(n_estimators=1, max_depth=2)
X, y = make_classification(10, n_features=4, random_state=42)
X = X[:, :2]
model.fit(X, y)
for cl in [None, 0.231, 1e-6, 0.9]:
if cl is not None:
model.init_.class_prior_ = np.array([cl, cl])
initial_types = [('input', FloatTensorType((None, X.shape[1])))]
model_onnx = convert_sklearn(model, initial_types=initial_types)
if "Regressor" in str(model_onnx):
raise AssertionError(str(model_onnx))
sess = InferenceSession(model_onnx.SerializeToString())
res = sess.run(None, {'input': X.astype(np.float32)})
pred = model.predict_proba(X)
delta = abs(res[1][0][0] - pred[0, 0])
if delta > 1e-5:
rows = ["diff", str(delta),
"X", str(X),
"base_values_", str(model.init_.class_prior_),
"predicted_label", str(model.predict(X)),
"expected", str(pred),
"onnxruntime", str(DataFrame(res[1])),
"model", str(model_onnx)]
raise AssertionError("\n---\n".join(rows))
dump_binary_classification(
model, suffix="1Deviance",
allow_failure="StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('%s')" % THRESHOLD)
def test_gradient_boosting_classifier3(self):
model = GradientBoostingClassifier(n_estimators=3)
dump_binary_classification(
model,
suffix="3",
allow_failure="StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.5.0')")
def test_random_forest_classifier(self):
model = RandomForestClassifier(n_estimators=3)
dump_one_class_classification(model,
allow_failure="StrictVersion(onnx.__version__) < StrictVersion('1.2')")
dump_binary_classification(model,
allow_failure="StrictVersion(onnx.__version__) < StrictVersion('1.2')")
dump_multiple_classification(model,
allow_failure="StrictVersion(onnx.__version__) < StrictVersion('1.2')")
def test_lightgbm_classifier(self):
model = LGBMClassifier(n_estimators=3, min_child_samples=1)
dump_binary_classification(
model, allow_failure="StrictVersion(onnx.__version__) < "
"StrictVersion('1.3.0')")
dump_multiple_classification(
model,
allow_failure="StrictVersion(onnx.__version__) < "
"StrictVersion('1.3.0')")
def test_lightgbm_classifier(self):
model = LGBMClassifier(n_estimators=3, min_child_samples=1)
dump_binary_classification(model,
target_opset={
'': TARGET_OPSET,
'ai.onnx.ml': TARGET_OPSET_ML
})
dump_multiple_classification(model,
target_opset={
'': TARGET_OPSET,
'ai.onnx.ml': TARGET_OPSET_ML
})
def test_xgb_classifier(self):
iris = load_iris()
X = iris.data[:, :2]
y = iris.target
y[y == 2] = 0
xgb = XGBClassifier(n_estimators=3)
xgb.fit(X, y)
conv_model = convert_sklearn(
xgb, initial_types=[
('input', FloatTensorType(shape=[None, X.shape[1]]))])
self.assertTrue(conv_model is not None)
dump_binary_classification(xgb, label_string=False)
def test_voting_soft_binary(self):
model = VotingClassifier(
voting="soft",
flatten_transform=False,
estimators=[
("lr", LogisticRegression()),
("lr2", LogisticRegression(fit_intercept=False)),
],
)
dump_binary_classification(model,
suffix="Soft",
comparable_outputs=[0, 1],
target_opset=TARGET_OPSET)
def test_xgb_classifier_reglog(self):
iris = load_iris()
X = iris.data[:, :2]
y = iris.target
y[y == 2] = 0
xgb = XGBClassifier(objective='reg:logistic')
xgb.fit(X, y)
conv_model = convert_sklearn(
xgb, initial_types=[
('input', FloatTensorType(shape=[None, X.shape[1]]))])
self.assertTrue(conv_model is not None)
dump_binary_classification(xgb, suffix="RegLog", label_string=False)
def test_voting_soft_binary_weighted(self):
model = VotingClassifier(
voting="soft",
flatten_transform=False,
weights=numpy.array([1.8, 0.2]),
estimators=[
("lr", LogisticRegression()),
("lr2", LogisticRegression(fit_intercept=False)),
],
)
dump_binary_classification(model,
suffix="WeightedSoft",
target_opset=TARGET_OPSET)
def test_voting_hard_binary(self):
model = VotingClassifier(
voting="hard",
flatten_transform=False,
estimators=[
("lr", LogisticRegression()),
("lr2", LogisticRegression(fit_intercept=False)),
],
)
# predict_proba is not defined when voting is hard.
dump_binary_classification(model,
suffix="Hard",
comparable_outputs=[0],
target_opset=TARGET_OPSET)
def test_voting_soft_binary(self):
model = VotingClassifier(voting='soft',
flatten_transform=False,
estimators=[
('lr', LogisticRegression()),
('lr2',
LogisticRegression(fit_intercept=False))
])
dump_binary_classification(
model,
suffix='Soft-OneOffArray',
comparable_outputs=[0, 1],
allow_failure=
"StrictVersion(onnxruntime.__version__) <= StrictVersion('0.2.1')")
def test_voting_soft_binary_weighted(self):
model = VotingClassifier(voting='soft',
flatten_transform=False,
weights=numpy.array([1.8, 0.2]),
estimators=[
('lr', LogisticRegression()),
('lr2',
LogisticRegression(fit_intercept=False))
])
dump_binary_classification(
model,
suffix='WeightedSoft-OneOffArray',
allow_failure=
"StrictVersion(onnxruntime.__version__) <= StrictVersion('0.2.1')")
def test_voting_hard_binary(self):
model = VotingClassifier(voting='hard',
flatten_transform=False,
estimators=[
('lr', LogisticRegression()),
('lr2',
LogisticRegression(fit_intercept=False))
])
# predict_proba is not defined when voting is hard.
dump_binary_classification(
model,
suffix='Hard-OneOffArray',
comparable_outputs=[0],
allow_failure=
"StrictVersion(onnxruntime.__version__) <= StrictVersion('0.2.1')")
def test_decision_tree_classifier(self):
model = DecisionTreeClassifier()
dump_one_class_classification(
model,
# Operator cast-1 is not implemented in onnxruntime
allow_failure=
"StrictVersion(onnx.__version__) < StrictVersion('1.2')")
dump_binary_classification(
model,
allow_failure=
"StrictVersion(onnx.__version__) < StrictVersion('1.2')")
dump_multiple_classification(
model,
allow_failure=
"StrictVersion(onnx.__version__) < StrictVersion('1.2')")
def test_voting_soft_binary_weighted(self):
model = VotingClassifier(
voting="soft",
flatten_transform=False,
weights=numpy.array([1.8, 0.2]),
estimators=[
("lr", LogisticRegression()),
("lr2", LogisticRegression(fit_intercept=False)),
],
)
dump_binary_classification(
model,
suffix="WeightedSoft",
allow_failure="StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')",
)
def test_voting_soft_binary(self):
model = VotingClassifier(
voting="soft",
flatten_transform=False,
estimators=[
("lr", LogisticRegression()),
("lr2", LogisticRegression(fit_intercept=False)),
],
)
dump_binary_classification(
model,
suffix="Soft",
comparable_outputs=[0, 1],
allow_failure="StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')",
)
def test_voting_hard_binary(self):
model = VotingClassifier(
voting="hard",
flatten_transform=False,
estimators=[
("lr", LogisticRegression()),
("lr2", LogisticRegression(fit_intercept=False)),
],
)
# predict_proba is not defined when voting is hard.
dump_binary_classification(
model,
suffix="Hard",
comparable_outputs=[0],
allow_failure="StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.5.0')",
)
def test_decision_tree_classifier(self):
model = DecisionTreeClassifier()
dump_one_class_classification(
model,
# Operator cast-1 is not implemented in onnxruntime
allow_failure="StrictVersion(onnx.__version__)"
" < StrictVersion('1.3') or "
"StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')",
)
dump_binary_classification(
model,
allow_failure="StrictVersion(onnx.__version__)"
" < StrictVersion('1.3') or "
"StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')",
)
dump_multiple_classification(
model,
allow_failure="StrictVersion(onnx.__version__)"
" < StrictVersion('1.3') or "
"StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')")
dump_multiple_classification(
model,
label_uint8=True,
allow_failure="StrictVersion(onnx.__version__)"
" < StrictVersion('1.3') or "
"StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')")
dump_multiple_classification(
model,
label_string=True,
allow_failure="StrictVersion(onnx.__version__)"
" < StrictVersion('1.3') or "
"StrictVersion(onnxruntime.__version__)"
" <= StrictVersion('0.2.1')")
def test_gradient_boosting_classifier(self):
model = GradientBoostingClassifier(n_estimators=3)
dump_binary_classification(model)
def test_extra_trees_classifier(self):
model = ExtraTreesClassifier(n_estimators=3)
dump_one_class_classification(model)
dump_binary_classification(model)
dump_multiple_classification(model)
def test_random_forest_classifier(self):
model = RandomForestClassifier(n_estimators=3)
dump_one_class_classification(model)
dump_binary_classification(model)
dump_multiple_classification(model)
def test_extra_tree_classifier(self):
model = ExtraTreeClassifier()
dump_one_class_classification(model)
dump_binary_classification(model)
dump_multiple_classification(model)
