Jimmy Wang

use `build/bin/FaceLandmarkImg -f sample1.jpg -aus`, or it will output other attributes so `prepare_au_annotations.py` cannot properly handle it

> ghm_loss.py", line 55, in _custom_loss_grad > return torch.sigmoid(x).detach() - target > RuntimeError: The size of tensor a (2) must match the size of tensor b (208) at non-singleton dimension...

It's when `(overlap> (torch.Tensor([0.6, 0.1, 0.2]) < 0.45).nonzero().squeeze(dim=1) tensor([1, 2]) >>> (torch.Tensor([0.1, 0.6]) < 0.45).nonzero().squeeze(dim=1) tensor([0]) >>> (torch.Tensor([0.6, 0.1]) < 0.45).nonzero().squeeze(dim=1) tensor([1]) >>> (torch.Tensor([0.1]) < 0.45).nonzero().squeeze(dim=1) tensor([0]) ```

应该是bin_count的移动平均的系数 我变量名基本上和原论文里是对应的