Open mujtabaasad opened 5 months ago
To resolve this issue, follow these steps:
git clone https://github.com/erikwijmans/Pointnet2_PyTorch.git
cd Pointnet2_PyTorch/pointnet2_ops_lib
pip install .
If you still encounter errors, it's likely due to your GCC version. Adjust the GCC version based on the specific error messages you receive.
Thanks @zhiqing0205 for reply!
To resolve this issue, follow these steps:
- Ensure your CUDA version is between 11.3 and 11.8.
- Install the PyTorch version that matches your CUDA version.
- Perform an offline installation:
git clone https://github.com/erikwijmans/Pointnet2_PyTorch.git cd Pointnet2_PyTorch/pointnet2_ops_lib pip install .
If you still encounter errors, it's likely due to your GCC version. Adjust the GCC version based on the specific error messages you receive.
thank you very much, I ran the model. But the result is about 93.7, which is worser than the results mentioned in the paper. What maybe the reason?
Hi @limaodaxia, have you pre-processed the dataset? There were two other similar issues solved by pre-processing the dataset from scratch (https://github.com/CVLAB-Unibo/crossmodal-feature-mapping/issues/6 and https://github.com/CVLAB-Unibo/crossmodal-feature-mapping/issues/3).
I am using the following pytorch implementation of pointnet. However, I am not able to install it. https://github.com/erikwijmans/Pointnet2_PyTorch How did you get it installed.
by the way did you used PointNet in the implementations?? as you mentioned PointMAE for pointcloud feature extraction