A simple C++ project demonstrating how to run inference using the ONNX Runtime.
This repo integrates compiled C++ code with machine learning models (.onnx, .pkl) and provides an interactive CLI for predictions.
- Loads ONNX models with ONNX Runtime C++ API
- Accepts user input via command line
- Runs inference and outputs predicted class
- Error handling for ONNX exceptions
- Cross‑language workflow (Python training → C++ inference)
pipelineCpp/ ┣ .vscode # VSCode settings ┣ hush.cpp / hush.exe # Example source & binary ┣ main.cpp / main.exe # Entry point for inference ┣ onnxruntime.dll # ONNX Runtime library ┣ Simple_model.onnx # Trained model (ONNX format) ┣ Simple_model.pkl # Model checkpoint (Python pickle) ┗ train_mode.py # Python script for training
- C++17 or later
- ONNX Runtime (C++ API)
- Python (for training scripts, optional)
-
Clone the repo:
-
git clone https://github.com/YourUsername/onnx-cpp-inference.git cd onnx-cpp-inference
-
Compile g++ main.cpp -o main.exe -I/path/to/onnxruntime/include -L/path/to/onnxruntime/lib -lonnxruntime
-
Run ./main.exe
- The model will classify the input and return the predicted class.
- Type -1 to exit.
Feel Free to fork thi s repo and intorduce your innovationn and include your own code. Happy codingg!!
