Pointnet_Pointnet2_pytorch

Pointnet_Pointnet2_pytorch copied to clipboard

no files in stanford_indoor3d

14

Train my own S3DIS dataset, no files are generated and errors are reported after running the code ‘Python collect_indoor3d_data.py’ D:\CCCCCCCC\Pointnet_Pointnet2_pytorch-master\Pointnet_Pointnet2_pytorch-master\data/s3dis/Stanford3dDataset_v1.2/Area_8/cg9_2/Annotations ERROR!! D:\CCCCCCCC\Pointnet_Pointnet2_pytorch-master\Pointnet_Pointnet2_pytorch-master\data/s3dis/Stanford3dDataset_v1.2/Area_8/cg9_3/Annotations D:\CCCCCCCC\Pointnet_Pointnet2_pytorch-master\Pointnet_Pointnet2_pytorch-master\data/s3dis/Stanford3dDataset_v1.2/Area_8/cg9_3/Annotations\close.txt D:\CCCCCCCC\Pointnet_Pointnet2_pytorch-master\Pointnet_Pointnet2_pytorch-master\data/s3dis/Stanford3dDataset_v1.2/Area_8/cg9_3/Annotations ERROR!! D:\CCCCCCCC\Pointnet_Pointnet2_pytorch-master\Pointnet_Pointnet2_pytorch-master\data/s3dis/Stanford3dDataset_v1.2/Area_8/cg9_4/Annotations D:\CCCCCCCC\Pointnet_Pointnet2_pytorch-master\Pointnet_Pointnet2_pytorch-master\data/s3dis/Stanford3dDataset_v1.2/Area_8/cg9_4/Annotations\close.txt D:\CCCCCCCC\Pointnet_Pointnet2_pytorch-master\Pointnet_Pointnet2_pytorch-master\data/s3dis/Stanford3dDataset_v1.2/Area_8/cg9_4/Annotations ERROR!! D:\CCCCCCCC\Pointnet_Pointnet2_pytorch-master\Pointnet_Pointnet2_pytorch-master\data/s3dis/Stanford3dDataset_v1.2/Area_8/cg9_5/Annotations...

chenjing3340

Computations involving variables that require gradients will keep history. avoid using such variables in computations which will live beyond your training loops. Reference to PyTorch FAQ: "https://github.com/lisroach/pytorch/blob/53fceff1e1e08e6d11663d260e136eaa1a74be79/docs/source/notes/faq.rst"

ZhipengXue97

Index out of bounds

3

https://github.com/yanx27/Pointnet_Pointnet2_pytorch/blob/eb64fe0b4c24055559cea26299cb485dcb43d8dd/models/pointnet2_utils.py#L102

vienngo

Insufficient part segmentation accuracy when point cloud information is incomplete

1

When I experiment with shapenet's full point cloud data, the segmentation accuracy is often as good as described in the paper. However, when I deployed the network to my project,...

Jinbo-Zuo

OJack-zhangO

T-Net

3

T-Net is used to guarantee the rotational invariance of pointnet, etc. But I can`t find T-Net in PointNetSetAbstraction or PointNetSetAbstractionMsg, anyone can tell where it is.

timelovery

关于FPS采样

1

大佬，最远点采样的覆盖范围更好，能够尽可能的包含点云的整体特征，但是经过sa1的时候，数据经过多个mlp从3维升到D维了，在sa2或者sa3的阶段，依旧采用fps进行下采样还适用吗？依旧采用FPS是不是可以理解为pointnet++默认了升维前点A与点B之间的距离最远，则升维后他们的距离依旧最远，请问他是如何做到这种“全局结构保持的”

fish-kong

Question regarding Dataloaders for Semantic Segmentation using PointNet++

1

Hi, I am slightly confused about a couple of things. 1. Why are there two separate dataloaders for train_semseg and test_semseg in S3DISDataLoader.py? Both seem to use different techniques for...

An-u-rag

pth 转 pt 时会失败

1

import torch from models import pointnet2_part_seg_msg def to_categorical(y, num_classes): """ 1-hot encodes a tensor """ new_y = torch.eye(num_classes)[y.cpu().data.numpy(),] if (y.is_cuda): return new_y.cuda() return new_y model = pointnet2_part_seg_msg.get_model(4, False) model.eval() z=torch.load('b2.pth')...

wangsale

CUDA out of memory while doing part segmentation training

2

I tried to run the command below: `python train_partseg.py --model pointnet2_part_seg_msg --normal --log_dir pointnet2_part_seg_msg` but I got cuda error. Here's the detail: ` PARAMETER ... Namespace(model='pointnet2_part_seg_msg', batch_size=16, epoch=251, learning_rate=0.001, gpu='0',...

blueeaglex