JP

> The detection performance of the mask-wearing image is good for the distant face, but in the case of the mask-wearing image in the front, that is, close-up (a composition...

> 直接用我提供的模型就可以 感谢回复，如果直接用你公开的yolo7s-face权值作为预训练模型，那训练参数还可以用默认值？

There is also a problem that the training is completed without generating the best accuracy pth file（best_accuracy_top1.pth）

Thanks for your reply. I have added the training data augmentation in this repo into insightFace project. And obtain the above experimental results.

感谢回复，我转onnx模型是准备部署在arm（瑞芯微率rv1126）上，由于后处理头操作太复杂，导致部分操作不支持，我比较困惑的是为什么我导出的onnx模型和你的差异这么多？

> > 感谢回复，我转onnx模型是准备部署在arm（瑞芯微率rv1126）上，由于后处理头操作太复杂，导致部分操作不支持，我比较困惑的是为什么我导出的onnx模型和你的差异这么多？ > >  好的，十分感谢