About the pretrained model.

Hi @markendo, thanks for the interesting work. May I ask that the provided pretrained models correspond to the five set of results shown in Table 2 of the main paper? It seems that I get different results from Table 2. May I also confirm that you use all the subsets of AMASS dataset? I am asking because I noticed that CNRS and WEIZMANN do not have the SMPL + H format.