sklearn SVM model converted into ONNX model but gives different output results

I am using sklearn to train a SVM model and I converted it into ONNX model. However, when testing the two models with the same input, the outputs are different.

System information

• OS Platform and Distribution: Windows 10 Enterprise 64 bit
• ONNX Runtime installed from (source or binary): install using PIP from Python 3.9.7
• ONNX Runtime version: 1.11.0
• sklearn version: 1.0.2
• Python version: 3.9.7
• Visual Studio version (if applicable): N/A
• GCC/Compiler version (if compiling from source): N/A
• CUDA/cuDNN version: N/A
• GPU model and memory: Not using GPU

To Reproduce

import pickle
from sklearn import svm
import numpy as np
from skl2onnx import convert_sklearn
from skl2onnx.common.data_types import FloatTensorType
import onnxruntime as rt
svm1_file = 'svm.model'
classifier_1 = pickle.load(open(svm1_file, 'rb'))
testingexample = [62, 62, 99, 99, 99, 99, 297, 297, 343, 343, 512, 0.0925, 262, 612, 322.6, 6.7963, 0.01993]
features = np.array([testingexample]).astype(np.float32)
result1_1 = classifier_1.predict(features)[0] #SKLEARN Result
result1_2 = classifier_1.predict_proba(features)[0]
print('svm1_file, SKLEARN result is %s, %s\n'%(str(result1_1), str(result1_2)))
 
#Convert into ONNX
initial_types = [('float_input',FloatTensorType([None, 17]))]
onx1 = convert_sklearn(classifier_1, initial_types = initial_types)
with open("Convert_svm.onnx", "wb") as f:
        f.write(onx1.SerializeToString())
sess1 = rt.InferenceSession(onx1.SerializeToString())
res1 = sess1.run(None, {'float_input': features.astype(np.float32)})
print(res1) #ONNX result
 

Sklearn result is 0.0, [0.99006934 0.00993066]
Onnx result is [array([0], dtype=int64), [{0: 0.9864753484725952, 1: 0.01352467481046915}]]

Expected behavior
The Sklearn model result should match the ONNX model result.
Onnx_and_sklearn_Models.zip

Is there anything wrong in my code when converting the SVM learned by using SKLEARN into ONNX format? Or, is this level of discrepancies expected?