Thanks so much for sharing the visualization code !!!
When using your visualization code, it works well on the GRAB and NTU datasets.
But on HumanAct12 and BABEL dataset, the visualization failed.
For HumanAct12, it shows
Traceback (most recent call last):
File "/home/WAT-main/eval_vae_act_render_video.py", line 202, in
val(args.iter)
File "/home/WAT-main/eval_vae_act_render_video.py", line 101, in val
render_videos_new(sequence, device, cfg.result_dir + f'/{args.mode}', key, w_golbalrot=True, smpl_model=smpl_model)
File "/home/WAT-main/utils/vis_util.py", line 372, in render_videos_new
verts_sbj = to_cpu(sbj_m(*sbj_parms).vertices)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(input, **kwargs)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/body_models.py", line 372, in forward
self.lbs_weights, pose2rot=pose2rot)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/lbs.py", line 220, in lbs
pose_feature, posedirs).view(batch_size, -1, 3)
RuntimeError: mat1 dim 1 must match mat2 dim 0
Process finished with exit code 1
and For BABEL, it shows
Traceback (most recent call last):
File "/home/WAT-main/eval_vae_act_render_video.py", line 202, in
val(args.iter)
File "/home/WAT-main/eval_vae_act_render_video.py", line 101, in val
render_videos_new(sequence, device, cfg.result_dir + f'/{args.mode}', key, w_golbalrot=True, smpl_model=smpl_model)
File "/home/WAT-main/utils/vis_util.py", line 372, in render_videos_new
verts_sbj = to_cpu(sbj_m(*sbj_parms).vertices)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(input, **kwargs)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/body_models.py", line 731, in forward
self.lbs_weights, pose2rot=pose2rot)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/lbs.py", line 230, in lbs
J_transformed, A = batch_rigid_transform(rot_mats, J, parents, dtype=dtype)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/lbs.py", line 380, in batch_rigid_transform
rel_joints.reshape(-1, 3, 1)).reshape(-1, joints.shape[1], 4, 4)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/lbs.py", line 341, in transform_mat
F.pad(t, [0, 0, 0, 1], value=1)], dim=2)
RuntimeError: Sizes of tensors must match except in dimension 0. Got 52 and 1560 (The offending index is 0)
Process finished with exit code 1
I wonder if you could kindly provide more clues about the visualization on these two datasets.
P.s. I followed all the directions of downloading the smpl/smplh/smplx models. I also followed https://github.com/vchoutas/smplx/blob/main/tools/README.md and merged SMPL-H and MANO parameters for visualizing hands.
Thanks so much for sharing the visualization code !!!
When using your visualization code, it works well on the GRAB and NTU datasets. But on HumanAct12 and BABEL dataset, the visualization failed.
For HumanAct12, it shows
Traceback (most recent call last): File "/home/WAT-main/eval_vae_act_render_video.py", line 202, in
val(args.iter)
File "/home/WAT-main/eval_vae_act_render_video.py", line 101, in val
render_videos_new(sequence, device, cfg.result_dir + f'/{args.mode}', key, w_golbalrot=True, smpl_model=smpl_model)
File "/home/WAT-main/utils/vis_util.py", line 372, in render_videos_new
verts_sbj = to_cpu(sbj_m(*sbj_parms).vertices)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(input, **kwargs)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/body_models.py", line 372, in forward
self.lbs_weights, pose2rot=pose2rot)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/lbs.py", line 220, in lbs
pose_feature, posedirs).view(batch_size, -1, 3)
RuntimeError: mat1 dim 1 must match mat2 dim 0
Process finished with exit code 1
and For BABEL, it shows
Traceback (most recent call last): File "/home/WAT-main/eval_vae_act_render_video.py", line 202, in
val(args.iter)
File "/home/WAT-main/eval_vae_act_render_video.py", line 101, in val
render_videos_new(sequence, device, cfg.result_dir + f'/{args.mode}', key, w_golbalrot=True, smpl_model=smpl_model)
File "/home/WAT-main/utils/vis_util.py", line 372, in render_videos_new
verts_sbj = to_cpu(sbj_m(*sbj_parms).vertices)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/torch/nn/modules/module.py", line 727, in _call_impl
result = self.forward(input, **kwargs)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/body_models.py", line 731, in forward
self.lbs_weights, pose2rot=pose2rot)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/lbs.py", line 230, in lbs
J_transformed, A = batch_rigid_transform(rot_mats, J, parents, dtype=dtype)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/lbs.py", line 380, in batch_rigid_transform
rel_joints.reshape(-1, 3, 1)).reshape(-1, joints.shape[1], 4, 4)
File "/home/anaconda3/envs/pytorchRtx/lib/python3.7/site-packages/smplx/lbs.py", line 341, in transform_mat
F.pad(t, [0, 0, 0, 1], value=1)], dim=2)
RuntimeError: Sizes of tensors must match except in dimension 0. Got 52 and 1560 (The offending index is 0)
Process finished with exit code 1
I wonder if you could kindly provide more clues about the visualization on these two datasets. P.s. I followed all the directions of downloading the smpl/smplh/smplx models. I also followed https://github.com/vchoutas/smplx/blob/main/tools/README.md and merged SMPL-H and MANO parameters for visualizing hands.
Sorry for disturbing you and thanks again!