Doudi

167 line of trainer/parsernet.py: ``` # current_argmax: [batch, 7702]; self.layer_neigh: [7702, 20] current_argmax = torch.argmax(weights_from_net, axis=2) idx = torch.stack([torch.index_select(self.layer_neigh, 1, current_argmax[i])[0] for i in range(self.bs)]) ``` we want to get...

About the patch partition.

1

It seems that the PatchEmbed module embeds images and divide it into pathes, which is implemented with dilated convolution and large kernel size. But how it realize the locality demonstrated...

Trouble in SMPL estimation from 3D scans

2

Hi, thansk for the awesome work. I have tried to fit SMPL to Thuman2.0 dataset, but it fails as the image shows. The pretrained model works well with the test...

Thanks for awesome work. It helps me to diffuse the anchor feature to $R^3$. When I try to use your method in my project with batchsize=3, it seems unsupported. `Line...