Z-Z-J
Z-Z-J
 In your paper：Specifically, for each frame we predict the body shape parameters. Then, we apply average pooling to get a single shape (βˆ) across the whole input sequence。 But...
Did you reproduce the accuracy reported in the paper
 Can you tell me more details about MeshNet's individual training?
Thank you for your work! Can you provide the SMPL annotations for Campus and Shelf datasets?
get_3dpw.py: it contains that how to process the original raw data and load processed data. But, get_mpi_inf.py: it is same as mpi-inf.py, only contains the code to load processed data....
Can you provide these data sets,? 1 .data/unfrequent_0.2_test_gt.npz 2. .data/data/mpi_inf_3dhp_small.npz 3. data/3dpw_{}_valid.npz