ONNX with FloatTensorType when inferred from C++ returns different label everytime

[VPU-tech]

My ONNX Model has been built on float pointing numbers, where number accuracy is of high importance upto 10 decimal precision .
ONNX model input is array of float numbers and output is label.

When I am inferring this ONNX model in C++ using 'Microsoft.ML.OnnxRuntime\1.13.1\runtimes', for same set of input, output label keeps differing. But when I infer same ONNX model from python, label output is consistent.

C++ code snippet:

---

int64_t* pShape = mInputDims.data();
size_t shape_len = mInputDims.size();

std::vector inputTensorValues(inputTensorSize);
inputTensorValues.assign(inputValues.begin(), inputValues.end()); //inputValues means std::vector

// Create input tensors of ORT::Value, which is a tensor format used by ONNX Runtime
Ort::MemoryInfo memoryInfo = Ort::MemoryInfo::CreateCpu(OrtAllocatorType::OrtArenaAllocator,
OrtMemType::OrtMemTypeDefault);
inputTensors.push_back(Ort::Value::CreateTensor(memoryInfo, inputTensorValues.data(),
inputTensorSize, pShape, shape_len));

// Create output tensor buffer
mOutputDims[0] = 1;
size_t outputTensorSize = vectorProduct(mOutputDims);
std::vector<int64_t> outputTensorValues(outputTensorSize);

// Create output tensors of ORT::Value
outputTensors.push_back(Ort::Value::CreateTensor<int64_t>(memoryInfo, outputTensorValues.data(), outputTensorSize,
mOutputDims.data(), mOutputDims.size()));

session->Run(Ort::RunOptions{ nullptr }, inputNames.data(), inputTensors.data(), 1,
outputNames.data(), outputTensors.data(), 1);

// Get the inference result from Labels list; count = LabelList's size
int64_t* outArr = outputTensors.front().GetTensorMutableData<int64_t>();
int64_t cls_idx = std::max_element(outArr, outArr + count) - outArr;
m_label = labels[cls_idx];

---

what can be the reason for this inconsistency of prediction?
In python float value is of 64 bit and in C++ float value is 8-bit. Can this data loss be the reason of inconsistency? How to deal with it?

Thanks in advance,
Best Regards,
Veena

[VPU-tech]

Iris_ONNX_Inference_Sample.zip

Code snippet attached here. ONNX model inferred from Python as well as from C++ module.
'iris_predictions_results_dump.csv' shows prediction results, generated using ONNX Inference from Python as well as from C++ Module.

Please help me understand why results from these 2 Inferences are not matching.

Thanks and Regards,
Veena