Issues with generating "reconstructed_poses.hdf5"

[treesturn]

I attempted to use the suggested videoavatars repo to generate the reconstructed_poses.hdf5 poses using the sample people_snapshot dataset: masks.hdf5, keypoints.hdf5 and camera.pkl, specifically the run_step1.sh file, but this is the error im getting, does anyone know how to get the 'reconstructed_poses.hdf5'

[kitylam9]

Need to use the SMPL v1.0.0 model file, basicmodel_f_lbs_10_207_0_v1.0.0.pkl However, even the model is correct, the training is not correct after some point, the error I got is Traceback (most recent call last): File "step1_pose.py", line 380, in args.body_height, args.nohands, args.display) File "step1_pose.py", line 322, in main fit_pose(current_frame, last_smpl, frustum, nohands, debug_rn) File "step1_pose.py", line 197, in fit_pose bgcolor=0, num_channels=1) File "/home/lambeyond/.local/lib/python2.7/site-packages/chumpy/ch.py", line 129, in new result.set(**defs) File "/home/lambeyond/.local/lib/python2.7/site-packages/chumpy/ch.py", line 482, in set self.setattr(kwall[-1], kwargs[kwall[-1]], None) File "/home/lambeyond/.local/lib/python2.7/site-packages/chumpy/ch.py", line 407, in setattr object.setattr(self, name, value) AttributeError: can't set attribute

[pengsida]

Please open an issue at the repository of videoavatars.