Probabilities or raw scores

A classifier usually returns a matrix of probabilities. By default, sklearn-onnx creates an ONNX graph which returns probabilities but it may skip that step and return raw scores if the model implements the method decision_function. Option 'raw_scores' is used to change the default behaviour. Let’s see that on a simple example.

Train a model and convert it

import numpy
import sklearn
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
import onnxruntime as rt
import onnx
import skl2onnx
from skl2onnx.common.data_types import FloatTensorType
from skl2onnx import convert_sklearn
from sklearn.linear_model import LogisticRegression

iris = load_iris()
X, y = iris.data, iris.target
X_train, X_test, y_train, y_test = train_test_split(X, y)
clr = LogisticRegression(max_iter=500)
clr.fit(X_train, y_train)
print(clr)

initial_type = [("float_input", FloatTensorType([None, 4]))]
onx = convert_sklearn(clr, initial_types=initial_type, target_opset=12)

LogisticRegression(max_iter=500)

Output type

Let’s confirm the output type of the probabilities is a list of dictionaries with onnxruntime.

sess = rt.InferenceSession(onx.SerializeToString(), providers=["CPUExecutionProvider"])
res = sess.run(None, {"float_input": X_test.astype(numpy.float32)})
print("skl", clr.predict_proba(X_test[:1]))
print("onnx", res[1][:2])

skl [[9.82794559e-01 1.72053489e-02 9.16403830e-08]]
onnx [{0: 0.9827945232391357, 1: 0.017205340787768364, 2: 9.164028114128087e-08}, {0: 0.00189912598580122, 1: 0.4566256105899811, 2: 0.541475236415863}]

Raw scores and decision_function

initial_type = [("float_input", FloatTensorType([None, 4]))]
options = {id(clr): {"raw_scores": True}}
onx2 = convert_sklearn(
    clr, initial_types=initial_type, options=options, target_opset=12
)

sess2 = rt.InferenceSession(
    onx2.SerializeToString(), providers=["CPUExecutionProvider"]
)
res2 = sess2.run(None, {"float_input": X_test.astype(numpy.float32)})
print("skl", clr.decision_function(X_test[:1]))
print("onnx", res2[1][:2])

skl [[ 6.74440614  2.69922635 -9.44363249]]
onnx [{0: 6.744406700134277, 1: 2.6992263793945312, 2: -9.443633079528809}, {0: -3.7117910385131836, 1: 1.770678997039795, 2: 1.9411125183105469}]

Versions used for this example

print("numpy:", numpy.__version__)
print("scikit-learn:", sklearn.__version__)
print("onnx: ", onnx.__version__)
print("onnxruntime: ", rt.__version__)
print("skl2onnx: ", skl2onnx.__version__)

numpy: 1.23.5
scikit-learn: 1.4.dev0
onnx:  1.15.0
onnxruntime:  1.16.0+cu118
skl2onnx:  1.15.0