Hi,

I'd like to know how the body shape (the beta parameters) is considered in the evaluations?
For the quantitative results in the paper, does the model uses default shape parameters (torch.zeros(B, num_shape)) or does it utilize the GT beta values?

Thank you very much