Note
Click here to download the full example code
Train, convert and predict with ONNX Runtime¶
This example demonstrates an end to end scenario starting with the training of a machine learned model to its use in its converted from.
Train a logistic regression¶
The first step consists in retrieving the iris datset.
from sklearn.datasets import load_iris
iris = load_iris()
X, y = iris.data, iris.target
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y)
Then we fit a model.
from sklearn.linear_model import LogisticRegression
clr = LogisticRegression()
clr.fit(X_train, y_train)
We compute the prediction on the test set and we show the confusion matrix.
from sklearn.metrics import confusion_matrix
pred = clr.predict(X_test)
print(confusion_matrix(y_test, pred))
[[12 0 0]
[ 0 8 0]
[ 0 3 15]]
Conversion to ONNX format¶
We use module sklearn-onnx to convert the model into ONNX format.
from skl2onnx import convert_sklearn
from skl2onnx.common.data_types import FloatTensorType
initial_type = [("float_input", FloatTensorType([None, 4]))]
onx = convert_sklearn(clr, initial_types=initial_type)
with open("logreg_iris.onnx", "wb") as f:
f.write(onx.SerializeToString())
We load the model with ONNX Runtime and look at its input and output.
import onnxruntime as rt
sess = rt.InferenceSession("logreg_iris.onnx", providers=rt.get_available_providers())
print("input name='{}' and shape={}".format(sess.get_inputs()[0].name, sess.get_inputs()[0].shape))
print("output name='{}' and shape={}".format(sess.get_outputs()[0].name, sess.get_outputs()[0].shape))
input name='float_input' and shape=[None, 4]
output name='output_label' and shape=[None]
We compute the predictions.
input_name = sess.get_inputs()[0].name
label_name = sess.get_outputs()[0].name
import numpy
pred_onx = sess.run([label_name], {input_name: X_test.astype(numpy.float32)})[0]
print(confusion_matrix(pred, pred_onx))
[[12 0 0]
[ 0 11 0]
[ 0 0 15]]
The prediction are perfectly identical.
Probabilities¶
Probabilities are needed to compute other relevant metrics such as the ROC Curve. Let’s see how to get them first with scikit-learn.
prob_sklearn = clr.predict_proba(X_test)
print(prob_sklearn[:3])
[[2.84438254e-05 2.76876851e-02 9.72283871e-01]
[3.93053588e-02 9.53960627e-01 6.73401437e-03]
[8.22028452e-02 9.05171919e-01 1.26252360e-02]]
And then with ONNX Runtime. The probabilies appear to be
prob_name = sess.get_outputs()[1].name
prob_rt = sess.run([prob_name], {input_name: X_test.astype(numpy.float32)})[0]
import pprint
pprint.pprint(prob_rt[0:3])
[{0: 2.844382106559351e-05, 1: 0.0276876799762249, 2: 0.9722838997840881},
{0: 0.039305318146944046, 1: 0.9539607167243958, 2: 0.006734013557434082},
{0: 0.08220282196998596, 1: 0.9051719903945923, 2: 0.012625232338905334}]
Let’s benchmark.
from timeit import Timer
def speed(inst, number=10, repeat=20):
timer = Timer(inst, globals=globals())
raw = numpy.array(timer.repeat(repeat, number=number))
ave = raw.sum() / len(raw) / number
mi, ma = raw.min() / number, raw.max() / number
print("Average %1.3g min=%1.3g max=%1.3g" % (ave, mi, ma))
return ave
print("Execution time for clr.predict")
speed("clr.predict(X_test)")
print("Execution time for ONNX Runtime")
speed("sess.run([label_name], {input_name: X_test.astype(numpy.float32)})[0]")
Execution time for clr.predict
Average 4.33e-05 min=4.17e-05 max=5.54e-05
Execution time for ONNX Runtime
Average 2.33e-05 min=2.27e-05 max=2.89e-05
2.325732499997457e-05
Let’s benchmark a scenario similar to what a webservice experiences: the model has to do one prediction at a time as opposed to a batch of prediction.
def loop(X_test, fct, n=None):
nrow = X_test.shape[0]
if n is None:
n = nrow
for i in range(0, n):
im = i % nrow
fct(X_test[im : im + 1])
print("Execution time for clr.predict")
speed("loop(X_test, clr.predict, 100)")
def sess_predict(x):
return sess.run([label_name], {input_name: x.astype(numpy.float32)})[0]
print("Execution time for sess_predict")
speed("loop(X_test, sess_predict, 100)")
Execution time for clr.predict
Average 0.00396 min=0.00394 max=0.00405
Execution time for sess_predict
Average 0.0011 min=0.00109 max=0.00112
0.0011013778150000065
Let’s do the same for the probabilities.
print("Execution time for predict_proba")
speed("loop(X_test, clr.predict_proba, 100)")
def sess_predict_proba(x):
return sess.run([prob_name], {input_name: x.astype(numpy.float32)})[0]
print("Execution time for sess_predict_proba")
speed("loop(X_test, sess_predict_proba, 100)")
Execution time for predict_proba
Average 0.00594 min=0.00582 max=0.00657
Execution time for sess_predict_proba
Average 0.00116 min=0.00114 max=0.00118
0.0011605704699996977
This second comparison is better as ONNX Runtime, in this experience, computes the label and the probabilities in every case.
Benchmark with RandomForest¶
We first train and save a model in ONNX format.
from sklearn.ensemble import RandomForestClassifier
rf = RandomForestClassifier()
rf.fit(X_train, y_train)
initial_type = [("float_input", FloatTensorType([1, 4]))]
onx = convert_sklearn(rf, initial_types=initial_type)
with open("rf_iris.onnx", "wb") as f:
f.write(onx.SerializeToString())
We compare.
sess = rt.InferenceSession("rf_iris.onnx", providers=rt.get_available_providers())
def sess_predict_proba_rf(x):
return sess.run([prob_name], {input_name: x.astype(numpy.float32)})[0]
print("Execution time for predict_proba")
speed("loop(X_test, rf.predict_proba, 100)")
print("Execution time for sess_predict_proba")
speed("loop(X_test, sess_predict_proba_rf, 100)")
Execution time for predict_proba
Average 0.654 min=0.652 max=0.659
Execution time for sess_predict_proba
Average 0.00139 min=0.00137 max=0.00144
0.0013873463899994932
Let’s see with different number of trees.
measures = []
for n_trees in range(5, 51, 5):
print(n_trees)
rf = RandomForestClassifier(n_estimators=n_trees)
rf.fit(X_train, y_train)
initial_type = [("float_input", FloatTensorType([1, 4]))]
onx = convert_sklearn(rf, initial_types=initial_type)
with open("rf_iris_%d.onnx" % n_trees, "wb") as f:
f.write(onx.SerializeToString())
sess = rt.InferenceSession("rf_iris_%d.onnx" % n_trees, providers=rt.get_available_providers())
def sess_predict_proba_loop(x):
return sess.run([prob_name], {input_name: x.astype(numpy.float32)})[0]
tsk = speed("loop(X_test, rf.predict_proba, 100)", number=5, repeat=5)
trt = speed("loop(X_test, sess_predict_proba_loop, 100)", number=5, repeat=5)
measures.append({"n_trees": n_trees, "sklearn": tsk, "rt": trt})
from pandas import DataFrame
df = DataFrame(measures)
ax = df.plot(x="n_trees", y="sklearn", label="scikit-learn", c="blue", logy=True)
df.plot(x="n_trees", y="rt", label="onnxruntime", ax=ax, c="green", logy=True)
ax.set_xlabel("Number of trees")
ax.set_ylabel("Prediction time (s)")
ax.set_title("Speed comparison between scikit-learn and ONNX Runtime\nFor a random forest on Iris dataset")
ax.legend()
5
Average 0.0542 min=0.0537 max=0.055
Average 0.00111 min=0.00109 max=0.00114
10
Average 0.0859 min=0.0857 max=0.0861
Average 0.00112 min=0.00111 max=0.00115
15
Average 0.117 min=0.117 max=0.118
Average 0.00115 min=0.00114 max=0.00117
20
Average 0.149 min=0.149 max=0.149
Average 0.00115 min=0.00114 max=0.00118
25
Average 0.182 min=0.181 max=0.183
Average 0.00117 min=0.00116 max=0.00119
30
Average 0.213 min=0.213 max=0.214
Average 0.00118 min=0.00117 max=0.0012
35
Average 0.244 min=0.244 max=0.245
Average 0.00118 min=0.00117 max=0.00121
40
Average 0.275 min=0.275 max=0.275
Average 0.00121 min=0.00119 max=0.00124
45
Average 0.307 min=0.307 max=0.307
Average 0.0012 min=0.00119 max=0.00124
50
Average 0.338 min=0.338 max=0.338
Average 0.00122 min=0.00121 max=0.00126
<matplotlib.legend.Legend object at 0x7f55eaeaa6e0>
Total running time of the script: ( 3 minutes 4.037 seconds)
