Error: module 'pytorch_utils' has no attribute 'SharedMLP'

[minhncsocial]

Hi Erik,
Thank you for sharing pointnet source.
When I run your code I get error:

Traceback (most recent call last): File "/home/minhnc-lab/WORKSPACES/Python/GRASP_DETECTION/PointNet/Pointnet2_PyTorch/train_cls.py", line 118, in <module> model = Pointnet(input_channels=0, num_classes=40, use_xyz=True) File "/home/minhnc-lab/WORKSPACES/Python/GRASP_DETECTION/PointNet/Pointnet2_PyTorch/models/pointnet2_msg_cls.py", line 62, in __init__ use_xyz=use_xyz File "/home/minhnc-lab/WORKSPACES/Python/GRASP_DETECTION/PointNet/Pointnet2_PyTorch/models/../utils/pointnet2_modules.py", line 107, in __init__ self.mlps.append(pt_utils.SharedMLP(mlp_spec, bn=bn)) AttributeError: module 'pytorch_utils' has no attribute 'SharedMLP'

I can not find out pytorch_utils module.
Could you share me that file?

Thank you very much.
Minhh

[erikwijmans]

pytorch_utils is in this repo: https://github.com/erikwijmans/etw_pytorch_utils. That repo is already added as a git-submodule. Looks like I didn't include the recommendation to clone with --recursive in this readme for this repo. You can checkout the submodule with git submodule update --init --recursive.

[minhncsocial]

Thank Erik very much.
I got your file and above issue was solved.

But I got another error when I run pytorch_utils: CUDA driver version is insufficient for CUDA runtime version
Have you ever met this error?
/home/minhnc-lab/WORKSPACES/Python/GRASP_DETECTION/PointNet/Pointnet2_PyTorch/models/../utils/pytorch_utils/pytorch_utils.py:104: UserWarning: nn.init.kaiming_normal is now deprecated in favor of nn.init.kaiming_normal_. init(conv_unit.weight) /home/minhnc-lab/WORKSPACES/Python/GRASP_DETECTION/PointNet/Pointnet2_PyTorch/models/../utils/pytorch_utils/pytorch_utils.py:53: UserWarning: nn.init.constant is now deprecated in favor of nn.init.constant_. nn.init.constant(self[0].weight, 1.0) /home/minhnc-lab/WORKSPACES/Python/GRASP_DETECTION/PointNet/Pointnet2_PyTorch/models/../utils/pytorch_utils/pytorch_utils.py:54: UserWarning: nn.init.constant is now deprecated in favor of nn.init.constant_. nn.init.constant(self[0].bias, 0) /home/minhnc-lab/WORKSPACES/Python/GRASP_DETECTION/PointNet/Pointnet2_PyTorch/models/../utils/pytorch_utils/pytorch_utils.py:252: UserWarning: nn.init.constant is now deprecated in favor of nn.init.constant_. nn.init.constant(fc.bias, 0) epochs: 0%| | 0/200 [00:00<?, ?it/s] train: 0%| | 0/615 [00:00<?, ?it/s]CUDA kernel failed : CUDA driver version is insufficient for CUDA runtime version

[erikwijmans]

I have not gotten that error in particular. It seems like an issue with your cuda install however.

[minhncsocial]

Thank you for your kind reply.
Because my environment is able to run pointnet pytorch v1. So maybe your environment is a little special.
Could you mind share me your .yml environment file?

[erikwijmans]

I have been able to build/run this with pytorch 3.0/3.1/4.0/4.1.

[minhncsocial]

Haha. Thank you. I already install pytorch 4.0.
Could you export your conda environment?