Go implementation of Arcface inference.
Based on Arcface-go,replace the onnxruntime-go&gocv repo.
- The onnx-format models used in the code is "buffalo_l" from insightface.
- ONNX Runtime-GO Visit this repo,learn how to deploy onnxruntime with your own requirement (like Windows/Linux/Macos|Cuda/Openvino/Rocm/Coreml...).In my repo,it has been provided the dependencies needed for DirectML acceleration on the Windows platform.If you need Hardware acceleration in your ouw platform.You must prepare lib files and edit the file . /arcface/arcface.go in line 62.
- GOCV is required, because some codes borrowed from gocv to implement EstimateAffinePartial2DWithParams().
The example is too simple, detect faces in the input image and retrieve features of the first face.
go run example.go
- path to "buffalo_l" should be corrected in
example.go.