/src/main/cpp/vnnjni.cpp:59: undefined reference to `VNN_Destroy_Face'

/src/main/cpp/vnnjni.cpp:1197: undefined reference to `VNN_Destroy_ObjCount'

./../src/main/cpp/vnnjni.cpp:1313: undefined reference to `VNN_Destroy_DocRect'

Can I know a specific version of the CMake tool and NDK version which is used at the time of compiling demos

1. update libraries on android and ios, fix bugs in face parser and stylizing.
2. update android demo, support video portrait segment, support image mode of disney face and 3d game face.
3. add license declaration to source files in demos of different platforms

1. fix bugs in face parser and stylizing applications.
2. update android demo, support video portrait segment, support image mode of disney face and 3d game.
3. add license declaration to source files in demos of different platforms.

1. upload libraries, demos on windows and macos.
2. update face_count[1.0.0].vnnmodel, qrcode_detection[1.0.0].vnnmodel, old models can not run on pc.
3. modify readme.

Hi..great app!! Would love to know how to get models into .vnnmodel format so i can use custom trained models..that would be a great extra..as any search into .vnnmodel yields a search result of..u.

public static class VNN_Image { public long ori_fmt; /* orientation format enum of img*/ public int pix_fmt; /* pixel format enum of img / public int mode_fmt; / mode format enum of detection, can use video/picture/default / public int width; //image width public int height; //image height public int channels; //image channel public byte[] data; //cpu data public float[] rect; public int[] texture; / if is cpu-backend, texture is NULL. But if is gpu-backend, texture may be a VNN_Texture(android|ios) or a CVPixelbuffer(ios). */ }