yfji
yfji
assign_box给出的regression label不是ymin,xmin,ymax,xmax吗？为什么在decode box的时候网络的regression又变成yoffset,xoffset,h,w呢
> For questions in part 3, > > 1. in every mini-batch, we put 50% pairs of same identity and another 50% of different identity. > 2. In discriminator, we...
> > > For questions in part 3, > > > > > > 1. in every mini-batch, we put 50% pairs of same identity and another 50% of different...
> HifiFace is mainly researched on image based swapping instead of video. So it is normal that the model performs not that perfect in video, as we did not tune...