Pointnet_Pointnet2_pytorch
Pointnet_Pointnet2_pytorch copied to clipboard
yanx27
PointNet and PointNet++ implemented by pytorch (pure python) and on ModelNet, ShapeNet and S3DIS.
Just for documentation: It works perfectly with Torch 2.1 and CUDA 11.8. ```python >>> torch.__version__ '2.1.2' >>> torch.version.cuda '11.8' ``` Shall I update the README with a PR?
I got a problme. RuntimeError: Error(s) in loading state_dict for get_model: size mismatch for sa1.conv_blocks.0.0.weight: copying a param with shape torch.Size([32, 6, 1, 1]) from checkpoint, the shape in current...
Hi! Thanks for sharing your code. I have a dataset in which the number of points is different in each class and also in different objects of the same class....
@yanx27 , Hello yan, I am facing a problem when I train the model the Instance accuracy and class accuracy are stuck on a fixed value. `Epoch 4 (17/200): 100%|█████████████████████████████████████████|...
Hi! I managed to run the code but I found no result file, ideally a file with point clouds and the label of classes. Could you please help me in...
Hi Xu Yan, In your github repo for semantic segmentation, how to preprocess the S3DIS data because when i used collecct_indoor3d_data.py nothing is saved in stanford_indoor3d. Can you please resolve...
In line 268, ``` for l in range(NUM_CLASSES): iou_per_class_str += 'class %s weight: %.3f, IoU: %.3f \n' % ( seg_label_to_cat[l] + ' ' * (14 - len(seg_label_to_cat[l])), labelweights[l - 1],...
2023-01-19 22:17:58,794 - Model - INFO - PARAMETER ... 2023-01-19 22:17:58,794 - Model - INFO - Namespace(batch_size=24, decay_rate=0.0001, epoch=200, gpu='0', learning_rate=0.001, log_dir='pointnet2_cls_ssg2', model='pointnet2_cls_ssg', num_category=40, num_point=1024, optimizer='Adam', process_data=False, use_cpu=False, use_normals=False, use_uniform_sample=False)...
Hello, May I ask what difference does it makes to not use --process_data and then use it? How does it accelerate training? Thank you
https://github.com/yanx27/Pointnet_Pointnet2_pytorch/blob/eb64fe0b4c24055559cea26299cb485dcb43d8dd/models/pointnet2_utils.py#L87 group_idx = torch.arange(N, dtype=torch.long).to(device).view(1, 1, N).repeat([B, S, 1]) sqrdists = square_distance(new_xyz, xyz) group_idx[sqrdists > radius ** 2] = N group_idx = group_idx.sort(dim=-1)[0][:, :, :nsample] The element of group_idx is...
