ExportInstructions.md

Export the model for the inference

Refer to:

• Object Detection
• Classification
• Instance Segmentation
• Optical Flow

For local export-plus-inference examples inside this repo, use docker_run_inference_e2e_example.sh. It provides preset-driven end-to-end flows, including OWLv2 with ONNX Runtime.

Note

The opencv-dnn module is configured to load ONNX models(not dynamic axis) and .weights(i.e. darknet format) for YOLOv4.

For detailed instructions on exporting each model, please refer to the linked documents above.