This is the onnxruntime inference code for AnimeGANv3. Official code: https://github.com/TachibanaYoshino/AnimeGANv3
It includes face detection (yolov5-face，or using any other model) to corp face for portrait stylization.
Download yolov5-s face detection onnx model to models path from https://pan.baidu.com/s/1gQuKiVkRiKB1aBCmghbRng (password:w1gr).
For example, portrait sketch. And for other style, just replace the onnx model.
python demo_onnx.py --useFace
It also works well for panoramic image. Without use face detection.
python demo_onnx.py
| input | useFace | panoramic image |
|---|---|---|
 |
 |
 |
 |
 |
 |