Convert a model with a reduced list of operators

Some runtime dedicated to onnx do not implement all the operators and a converted model may not run if one of them is missing from the list of available operators. Some converters may convert a model in different ways if the users wants to blacklist some operators.

GaussianMixture

The first converter to change its behaviour depending on a black list of operators is for model GaussianMixture.

import onnxruntime
import onnx
import numpy
import os
from timeit import timeit
import numpy as np
import matplotlib.pyplot as plt
from onnx.tools.net_drawer import GetPydotGraph, GetOpNodeProducer
from onnxruntime import InferenceSession
from sklearn.mixture import GaussianMixture
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from skl2onnx import to_onnx

data = load_iris()
X_train, X_test = train_test_split(data.data)
model = GaussianMixture()
model.fit(X_train)

GaussianMixture()

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

Default conversion

model_onnx = to_onnx(
    model,
    X_train[:1].astype(np.float32),
    options={id(model): {"score_samples": True}},
    target_opset=12,
)
sess = InferenceSession(
    model_onnx.SerializeToString(), providers=["CPUExecutionProvider"]
)

xt = X_test[:5].astype(np.float32)
print(model.score_samples(xt))
print(sess.run(None, {"X": xt})[2])

[-1.55507181 -1.88300778 -3.57222331 -1.67405519 -1.85984688]
[[-1.5550716]
 [-1.8830082]
 [-3.572221 ]
 [-1.6740558]
 [-1.8598464]]

Display the ONNX graph.

pydot_graph = GetPydotGraph(
    model_onnx.graph,
    name=model_onnx.graph.name,
    rankdir="TB",
    node_producer=GetOpNodeProducer(
        "docstring", color="yellow", fillcolor="yellow", style="filled"
    ),
)
pydot_graph.write_dot("mixture.dot")

os.system("dot -O -Gdpi=300 -Tpng mixture.dot")

image = plt.imread("mixture.dot.png")
fig, ax = plt.subplots(figsize=(40, 20))
ax.imshow(image)
ax.axis("off")

(-0.5, 4796.5, 8425.5, -0.5)

Conversion without ReduceLogSumExp

Parameter black_op is used to tell the converter not to use this operator. Let’s see what the converter produces in that case.

model_onnx2 = to_onnx(
    model,
    X_train[:1].astype(np.float32),
    options={id(model): {"score_samples": True}},
    black_op={"ReduceLogSumExp"},
    target_opset=12,
)
sess2 = InferenceSession(
    model_onnx2.SerializeToString(), providers=["CPUExecutionProvider"]
)

xt = X_test[:5].astype(np.float32)
print(model.score_samples(xt))
print(sess2.run(None, {"X": xt})[2])

[-1.55507181 -1.88300778 -3.57222331 -1.67405519 -1.85984688]
[[-1.5550716]
 [-1.8830082]
 [-3.5722215]
 [-1.6740558]
 [-1.8598464]]

Display the ONNX graph.

pydot_graph = GetPydotGraph(
    model_onnx2.graph,
    name=model_onnx2.graph.name,
    rankdir="TB",
    node_producer=GetOpNodeProducer(
        "docstring", color="yellow", fillcolor="yellow", style="filled"
    ),
)
pydot_graph.write_dot("mixture2.dot")

os.system("dot -O -Gdpi=300 -Tpng mixture2.dot")

image = plt.imread("mixture2.dot.png")
fig, ax = plt.subplots(figsize=(40, 20))
ax.imshow(image)
ax.axis("off")

(-0.5, 4921.5, 13264.5, -0.5)

Processing time

print(
    timeit(
        stmt="sess.run(None, {'X': xt})", number=10000, globals={"sess": sess, "xt": xt}
    )
)

print(
    timeit(
        stmt="sess2.run(None, {'X': xt})",
        number=10000,
        globals={"sess2": sess2, "xt": xt},
    )
)

0.4618227000000843
0.5505055000000993

The model using ReduceLogSumExp is much faster.

If the converter cannot convert without…

Many converters do not consider the white and black lists of operators. If a converter fails to convert without using a blacklisted operator (or only whitelisted operators), skl2onnx raises an error.

try:
    to_onnx(
        model,
        X_train[:1].astype(np.float32),
        options={id(model): {"score_samples": True}},
        black_op={"ReduceLogSumExp", "Add"},
        target_opset=12,
    )
except RuntimeError as e:
    print("Error:", e)

Error: Operator 'Add' is black listed.

Versions used for this example

import sklearn  # noqa

print("numpy:", numpy.__version__)
print("scikit-learn:", sklearn.__version__)
import skl2onnx  # noqa

print("onnx: ", onnx.__version__)
print("onnxruntime: ", onnxruntime.__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