microsoft/onnxruntime-extensions
Publicmirrored fromhttps://github.com/microsoft/onnxruntime-extensionsAvailable
docs/new_operator.md
27lines · modecode
| 1 | # Add a Custom Operator in ONNXRuntime-Extensions |
| 2 | |
| 3 | Before implement a custom operator, you get the ONNX model with one or more ORT custom operators, created by ONNX converters, [ONNX-Script](https://github.com/microsoft/onnx-script), or [ONNX model API](https://onnx.ai/onnx/api/helper.html) and etc.. |
| 4 | |
| 5 | |
| 6 | ## 1. Generate the C++ template code of the Custom operator from the ONNX Model (optional) |
| 7 | python -m onnxruntime-extensions.cmd --cpp-gen <model_path> <repository_dir>` |
| 8 | If you are familiar with the ONNX model detail, you create the custom operator C++ classes directly. |
| 9 | |
| 10 | |
| 11 | ## 2. Implement the CustomOp Kernel Compute method in the generated C++ files. |
| 12 | the custom operator kernel C++ code exmaple can be found [operators](../operators/) folder, like [KernelGaussianBlur](../operators/cv2/gaussian_blur.hpp). All C++ APIs that can be used in the kernel implmentation are listed below |
| 13 | |
| 14 | * [ONNXRuntime Custom API docs](https://onnxruntime.ai/docs/api/c/struct_ort_custom_op.html) |
| 15 | * the third libraries API docs intergrated in ONNXRuntime Extensions the can be used in C++ code |
| 16 | - OpenCV API docs https://docs.opencv.org/4.x/ |
| 17 | - Google SentencePiece Library docs https://github.com/google/sentencepiece/blob/master/doc/api.md |
| 18 | - dlib(matrix and ML library) C++ API docs http://dlib.net/algorithms.html |
| 19 | - BlingFire Library https://github.com/microsoft/BlingFire |
| 20 | - Google RE2 Library https://github.com/google/re2/wiki/CplusplusAPI |
| 21 | - JSON library https://json.nlohmann.me/api/basic_json/ |
| 22 | |
| 23 | ## 3. Build and Test |
| 24 | - The unit tests can be implemented as Python or C++, check [test](../test) folder for more examples |
| 25 | - Check [build-package](./development.md) on how to build the different langauage package to be used for production. |
| 26 | |
| 27 | Please check the [contribution](../README.md#contributing) to see if it is possible to contribute the custom operator to onnxruntime-extensions. |
| 28 | |