Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fixed GradientBoostingRegression model conversion failure with init=zero #164

Merged
merged 9 commits into from
Jun 17, 2019
Merged

Fixed GradientBoostingRegression model conversion failure with init=zero #164

Changes from 1 commit
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
f1586d5
Fixed GradientBoostingRegression model conversion failure with init=zero
Jun 4, 2019
e9bcae9
Merge branch 'master' into prroy/gradient_boosting_regressor_bug
Jun 7, 2019
a507784
Merge branch 'master' into prroy/gradient_boosting_regressor_bug
vinitra Jun 10, 2019
6643072
Loop on fit n_estimators_ instead of passed n_estimators
Jun 12, 2019
b6a730f
Merge branch 'prroy/gradient_boosting_regressor_bug' of https://githu…
Jun 12, 2019
c4ad6c1
Added corresponding unit test
Jun 17, 2019
dba7fdf
Merge branch 'master' into prroy/gradient_boosting_regressor_bug
Jun 17, 2019
49d9df2
Fix error in scikit < 0.21
Jun 17, 2019
561e3a8
Merge branch 'master' into prroy/gradient_boosting_regressor_bug
Jun 17, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
15 changes: 9 additions & 6 deletions skl2onnx/operator_converters/gradient_boosting.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,13 +87,16 @@ def convert_sklearn_gradient_boosting_regressor(scope, operator, container):
attrs['name'] = scope.get_unique_operator_name(op_type)
attrs['n_targets'] = 1

# constant_ was introduced in scikit-learn 0.21.
if hasattr(op.init_, 'constant_'):
cst = [float(x) for x in op.init_.constant_]
elif op.loss == 'ls':
cst = [op.init_.mean]
if op.init_ == 'zero':
cst = np.zeros((operator.inputs[0].type.shape[0], op.loss_.K))
else:
cst = [op.init_.quantile]
# constant_ was introduced in scikit-learn 0.21.
if hasattr(op.init_, 'constant_'):
cst = [float(x) for x in op.init_.constant_]
elif op.loss == 'ls':
cst = [op.init_.mean]
else:
cst = [op.init_.quantile]
attrs['base_values'] = [float(x) for x in cst]

tree_weight = op.learning_rate
Expand Down