longbowzhang

@kajalsanklecha Hi, I am facing exactly the same problem as you, do you have any idea to solve it? Thanks in advance.

Similar concerns here. Quite curious why not apply deformation to the view directions? Could you explain a little bit? @keunhong Looking forward to you reply

Hi @keunhong Thank you so much for your sharing! Now I can reproduce similar results with my own dataset by (1) use a much lower lr (e.g., 1e-4) (2) set...

I think I fail to express myself clearly. > The normalizaiton is only applied on 2d input and bone_length. # Yes I got this! Let me use an extreme case...

Hi @Shimingyi Nice work! But I have two additional questions about the joint rotation discriminator which require your further explanation if possible. 1. The quaternions output by the network is...

Hi @Shimingyi Just a follow-up comment, as you have mentioned in the paper > our discriminator judges the realism of temporal sequences of _angular velocities_. However, I think the finite...

Hi @Shimingyi, Thanks a lot for your fast reply. I am also curious about the motivation of the Adversarial Rotation Loss section. > Due to the fact that the T-poses...

Hi @chengzhag, Sorry to bother you. I am also trying to reproduce the results presented in the paper. Now I am working on the first stage, which is training LEN...

Hello @litingsjj, Did you figure out the motivation behind the usage of ignore_ids?

> "In all we train K + 2 discriminators" K means one discriminator for each of the joints, 1 discriminator for the overall pose, and 1 discriminator for shape.