Add C++ ONNX Demo

[sctrueew]

This PR introduces a C++ demo for the RF-DETR model, allowing users to perform real-time object detection using an ONNX model. The demo supports various input sources, including images, videos, and live camera streams, with optional CUDA acceleration.

Key Features:
✅ Loads an RF-DETR model in ONNX format
✅ Supports image, video, and live camera inference
✅ Enables CPU and CUDA (GPU) execution
✅ Configurable confidence threshold for detections
✅ Outputs annotated images/videos with detected objects
✅ Uses COCO class labels for object recognition

Run Examples:
🔹 Image Inference:
Detect objects in a static image and save the output:

./main--model path/to/model.onnx --source_type image \
  --input path/to/image.jpg --output path/to/output.jpg \
  --conf 0.6 --labels path/to/coco.names

🔹 Video Inference:
Process a video file and save the annotated output:

./main --model path/to/model.onnx --source_type video \
  --input path/to/video.mp4 --output path/to/output.mp4 \
  --conf 0.5 --use_cuda

🔹 Live Camera Inference (Default ID 0):
Run inference on the default webcam (ID 0) with GPU acceleration:

./main --model path/to/model.onnx --source_type camera \
  --input 0 --conf 0.55 --use_cuda

🔹 Live Camera Inference (Specific Camera ID 1):
Run inference on a specific camera (ID 1):

./main --model path/to/model.onnx --source_type camera \
  --input 1 --conf 0.55

🔹 Get Help & Available Options:

./main --help
Dependencies:

• OpenCV (for image and video processing)
• ONNX Runtime (for model inference)
• CUDA (optional, for GPU acceleration)

[CLAassistant]

All committers have signed the CLA.

[probicheaux]

Hey @sctrueew, really cool work! Would love to merge. Can you sign the CLA?

[sctrueew]

Hey @sctrueew, really cool work! Would love to merge. Can you sign the CLA?

Yes, I have agreed the CLA

[SkalskiP]

Hi @sctrueew, thanks for accepting the CLA! Would it be possible for you to add a README in inference/cpp that explains how to set up the project, install the necessary dependencies, and run the model in C++? This would make it much easier for new users to get started with your demo. Thanks again for your contribution!

[sctrueew]

Hi @sctrueew, thanks for accepting the CLA! Would it be possible for you to add a README in inference/cpp that explains how to set up the project, install the necessary dependencies, and run the model in C++? This would make it much easier for new users to get started with your demo. Thanks again for your contribution!

Hi @SkalskiP, I have added a README file to the project. Please let me know if you need any further changes or additional information.

[isaacrob]

what's the utility of having a large scale devoted c++ demo in this otherwise python-only repo?

[Borda]

@sctrueew, thank you for your suggestion and for sending this PR I believe this repo is more about research, training, and experimentation, and your C++ demo, however, it is cool and I like it more on deployment, from which we try to solve with our other package inference, which is also optimized for this 🦝