I do fitting on the quantitative dataset, but when i run:
python ./prox/main.py --config cfg_files/RGB.yaml --recording_dir /export/home/lg/data/prox/quantitative/recordings/vicon_03301_01 --output_folder ./RGB_results --vposer_ckpt /export/home/lg/data/prox/quantitative/models/vposer_v1_0/ --part_segm_fn /export/home/lg/data/prox/quantitative/models/smplx_parts_segm.pkl --model_folder /export/home/lg/data/prox/quantitative/models,
The output code:
Processing: /export/home/lg/data/prox/quantitative/recordings/vicon_03301_01/Color/s001_frame_00001__00.00.00.023.jpg Found Trained Model: /export/home/lg/data/prox/quantitative/models/vposer_v1_0/snapshots/TR00_E096.pt /export/home/lg/.local/opt/anaconda3/envs/torch10/lib/python3.6/site-packages/smplx/body_models.py:280: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). param[:] = torch.tensor(params_dict[param_name]) Camera initialization done after 16.8501 Camera initialization final loss 721.0235 Stage 000 done after 33.7366 seconds Stage 001 done after 5.8872 seconds Stage 002 done after 65.9812 seconds Orientation: 0%| | 0/1 [01:45<?, ?it/s[1] 6203 segmentation fault (core dumped) python ./prox/main.py --config cfg_files/RGB.yaml --recording_dir | 3/5 [01:45<01:19, 39.56s/it]
What happened and how to solve this problem?
My environment:
CentOS7
GCC 7.3.0
python 3.6.13
torch 1.1.0 ( the version of torch in the requirements is torch==1.0.1.post2, but the human_body_prior package needs torch>=1.1)
I do fitting on the quantitative dataset, but when i run:
python ./prox/main.py --config cfg_files/RGB.yaml --recording_dir /export/home/lg/data/prox/quantitative/recordings/vicon_03301_01 --output_folder ./RGB_results --vposer_ckpt /export/home/lg/data/prox/quantitative/models/vposer_v1_0/ --part_segm_fn /export/home/lg/data/prox/quantitative/models/smplx_parts_segm.pkl --model_folder /export/home/lg/data/prox/quantitative/models, The output code:Processing: /export/home/lg/data/prox/quantitative/recordings/vicon_03301_01/Color/s001_frame_00001__00.00.00.023.jpg Found Trained Model: /export/home/lg/data/prox/quantitative/models/vposer_v1_0/snapshots/TR00_E096.pt /export/home/lg/.local/opt/anaconda3/envs/torch10/lib/python3.6/site-packages/smplx/body_models.py:280: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). param[:] = torch.tensor(params_dict[param_name]) Camera initialization done after 16.8501 Camera initialization final loss 721.0235 Stage 000 done after 33.7366 seconds Stage 001 done after 5.8872 seconds Stage 002 done after 65.9812 seconds Orientation: 0%| | 0/1 [01:45<?, ?it/s[1] 6203 segmentation fault (core dumped) python ./prox/main.py --config cfg_files/RGB.yaml --recording_dir | 3/5 [01:45<01:19, 39.56s/it]What happened and how to solve this problem? My environment: CentOS7 GCC 7.3.0 python 3.6.13 torch 1.1.0 ( the version of torch in the requirements is torch==1.0.1.post2, but the human_body_prior package needs torch>=1.1)