Implementation:Microsoft Onnxruntime InferenceSession Run For Validation

Metadata

Field	Value
Implementation Name	InferenceSession_Run_For_Validation
Repository	Microsoft_Onnxruntime
Source Repository	https://github.com/microsoft/onnxruntime
Type	API Doc
Language	Python
Domain	ML_Inference, Model_Conversion
Last Updated	2026-02-10
Workflow	Train_Convert_Predict
Pair	4 of 5

Overview

API documentation for using onnxruntime.InferenceSession to validate that an ONNX-converted model produces predictions matching the original scikit-learn model.

API Signature

onnxruntime.InferenceSession(onnx_path).run([label_name], {input_name: X_test})

Compared against:

model.predict(X_test)

Import

import onnxruntime as rt
import numpy
from sklearn.metrics import confusion_matrix

Code Reference

Reference	Location
Validation example	docs/python/examples/plot_train_convert_predict.py:L66-80

I/O Contract

Inputs

Parameter	Type	Required	Description
ONNX model file	`str` (path)	Yes	Path to the ONNX model file produced by the conversion step.
Test dataset	`numpy.ndarray`	Yes	The same test data used to validate the original scikit-learn model, cast to `numpy.float32`.
Original predictions	`numpy.ndarray`	Yes	Predictions from the original scikit-learn model for comparison.

Outputs

Output	Type	Description
ONNX predictions	`numpy.ndarray`	Predictions from the ONNX model via `session.run()`.
Confusion matrix	`numpy.ndarray`	Comparison matrix confirming predictions match (should be diagonal).

Usage Example

import onnxruntime as rt
import numpy
from sklearn.metrics import confusion_matrix

# Load the converted ONNX model
sess = rt.InferenceSession("logreg_iris.onnx", providers=rt.get_available_providers())

# Get input/output names
input_name = sess.get_inputs()[0].name
label_name = sess.get_outputs()[0].name

# Run inference with ONNX Runtime
pred_onx = sess.run([label_name], {input_name: X_test.astype(numpy.float32)})[0]

# Compare with original sklearn predictions
print(confusion_matrix(pred, pred_onx))

From the source at docs/python/examples/plot_train_convert_predict.py:L66-80:

import onnxruntime as rt

sess = rt.InferenceSession("logreg_iris.onnx", providers=rt.get_available_providers())

print(f"input name='{sess.get_inputs()[0].name}' and shape={sess.get_inputs()[0].shape}")
print(f"output name='{sess.get_outputs()[0].name}' and shape={sess.get_outputs()[0].shape}")

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))

A successful validation produces a diagonal confusion matrix, confirming that every prediction from the ONNX model matches the original scikit-learn prediction.

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment