hello ,when I excute the commad: ./train_val_modelnet.sh -g 0 -x modelnet_x3_l4 it just print "Train/Val with setting modelnet_x3_l4 on GPU 0!",but no any other action,why???

when I compile tf_sampling_so.so file some warning happened: nvcc warning : The 'compute_20', 'sm_20', and 'sm_21' architectures are deprecated, and may be removed in a future release (Use -Wno-deprecated-gpu-targets to suppress warning). but the tf_sampling_so.so compiled successfully then I run the command: ./train_val_shapenet.sh -g 0 -x shapenet_x8_2048_fps error messages in pointcnn_seg_shapenet_x8_2048_fps.txt like that: Traceback (most recent call last): File "../train_val_seg.py", line 295, in main() File "../train_val_seg.py", line 127, in main net = model.Net(points_augmented, features_augmented, is_training, setting) File "/home/whf/ZYM/PointCNN/pointcnn_seg.py", line 11, in init PointCNN.init(self, points, features, is_training, setting) File "/home/whf/ZYM/PointCNN/pointcnn.py", line 64, in init from sampling import tf_sampling File "/home/whf/ZYM/PointCNN/sampling/tf_sampling.py", line 15, in sampling_module=tf.load_op_library(os.path.join(BASE_DIR, 'tf_sampling_so.so')) File "/home/whf/anaconda3/envs/tf_gpu/lib/python3.6/site-packages/tensorflow/python/framework/load_library.py", line 58, in load_op_library lib_handle = py_tf.TF_LoadLibrary(library_filename, status) File "/home/whf/anaconda3/envs/tf_gpu/lib/python3.6/site-packages/tensorflow/python/framework/errors_impl.py", line 516, in exit c_api.TF_GetCode(self.status.status)) tensorflow.python.framework.errors_impl.NotFoundError: /home/whf/ZYM/PointCNN/sampling/tf_sampling_so.so: undefined symbol: _ZN10tensorflow8internal21CheckOpMessageBuilder9NewStringEv

My environment as follows： gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.4) tensorflow 1.6.0

Hello, thanks for sharing your great work with us here! I just downloaded the ScanNet dataset for semantic segmentation task, and I have finished the training and testing. When I wanted to visualize the results, I found that every ply file is just a tiny part of the whole scene. How can I visualize the result of the whole scene? I have no idea which files belong to the same scene. Does every h5 file contain several different scenes?

When I test the semantic3d data set, my computer has 64GB of memory. Memory Error appears when I prepare data and train. How much memory does it need?

Hi, I followed the steps in the Semantic3d dataset and used a custom dataset to train. I was able to create .h5 and all steps were successful. But when I run, ./train_val_semantic3d.sh -g 0 -x semantic3d_x4_2048_fps :
inside models/seg -> the log file shows the following error: tf_sampling_so.so: cannot open shared object file: No such file or directory

I checked the existing issues (https://github.com/charlesq34/pointnet2/issues/48) and made changes to Pointcnn/sampling/tf_sampling_compiler.sh but still did not work.

I am using the TensorFlow version = 1.15, python 3.6, conda environment(Used pip command to install tf as mentioned in one of the issues. Still didn't work) Any help on how to resolve this issue? Regards Niranjan

https://github.com/yangyanli/PointCNN/blob/master/pointnetpp_cls/utils/pointnet_util.py#L49

flake8 testing of https://github.com/yangyanli/PointCNN

$ flake8 . --count --select=E901,E999,F821,F822,F823 --show-source --statistics

./pointnetpp_cls/utils/pointnet_util.py:49:23: F821 undefined name 'tnet'
        grouped_xyz = tnet(grouped_xyz, tnet_spec)
                      ^
1     F821 undefined name 'tnet'

Hi, Thank you for sharing your work. My 3D face dataset contains large ply files. For example:

ply
format ascii 1.0
element vertex 166368
property float x
property float y
property float z
element face 156797
property list uchar int vertex_indices
end_header
-50.8063 31.0753 83.1526
...
3 37583 37611 37610 
...

And i should do the classification basing on these ply files but I have no idea how to implement it using PointCNN. What should i do?

Thank you. tsly

How did you fix the benchmark files to reach 9,305/2,606 instances, as you mentioned in issue #29 ?

Can you update the benchmark files with the fixed files (maybe scannet-labels.combined.tsv and classes_ObjClassification-ShapeNetCore55.txt in ./data_conversions)? Thanks.

Hi, I am able to compile the sampling file but can't use it. I am using python 2.7 and Tensorflow 1.7.0 When I want to import sampling in pointcnn.py, I got the following error :

  File "/home/yjm/Project/PointCNN11_24_new/code/pointcnn.py", line 69, in __init__
    from sampling import tf_sampling
ImportError: No module named sampling

However, when I compile sampling, I got no error. My tf_sampling_compile.sh file is :

#/bin/bash
PYTHON=python
nvcc=/usr/local/cuda/bin/nvcc
cudalib=/usr/local/cuda/lib64
tensorflow=/home/yjm/anaconda3/envs/py27_1/lib/python2.7/site-packages/tensorflow/include
TF_INC=$($PYTHON -c 'import tensorflow as tf;print(tf.sysconfig.get_include())')
TF_LIB=$($PYTHON -c 'import tensorflow as tf;print(tf.sysconfig.get_lib())')

$nvcc tf_sampling_g.cu -o tf_sampling_g.cu.o -c -O2 -DGOOGLE_CUDA=1 -x cu -Xcompiler -fPIC
g++ -std=c++11 tf_sampling.cpp tf_sampling_g.cu.o -o tf_sampling_so.so -shared -fPIC -I $tensorflow -I$TF_INC/external/nsync/public -L$TF_LIB -ltensorflow_framework -I /usr/local/cuda/include -lcudart -L /usr/local/cuda/lib64/ -O2

I would be very grateful if you give me some suggestions ! Thanks!

Hi,

I tried to reproduce the results for Scannet Dataset on Segmentation Task but I cannot reproduce the mentioned results. The accuracy just around 77.76% (Validation) and ~85% for training data. I used the "pointnet++ preprocessed data" with provided scannet-seg hyper-parameters.

Can you give me a clue what I should do to reproduce the result? And also how long do you train the model to achieve the published result?

Hi, I could not run the code using the commands you give.

(pointcnn) [email protected]:/mnt/Ubuntu/PointCNN-master/pointcnn_cls$ ./train_val_modelnet.sh -g 0 -x modelnet_x3_l4 Train/Val with setting modelnet_x3_l4 on GPU 0! (pointcnn) [email protected]:/mnt/Ubuntu/PointCNN-master/pointcnn_cls$

It echos "Train/Val with setting modelnet_x3_l4 on GPU 0!" but exits immediately. What's the problem?

Thanks for your work! When i use your code extract scannet dataset for 3D classification , I get 12060 examples for training and 3416 for testing. It seems that scannetv2 is different from scannetv1. However , when i train it with pointnet and pointnet++, I respectively get the best performance 89.34 and 90.35! It is higher than your paper's result. Do anyone have the same founding with me? Is it normal phenomenon due to the Scannetv2 is easy to classify than Scannetv1(i guess that in 2018 , Scannet just release V1) or i make some mistakes which cause my dataset wrong?

Thanks for anyone's answer. Best regards. ╰(°▽°)╯

Dear author,

I am interested in finding out the model size for PointCNN part segmentation, but I could only find the information for PointCNN classification network. Could you please report what is GMac for PointCNN part segmentation network? Thank you very much in advance.

Hi. Thank you for the repository.

I wanted to ask if it is possible to get any guidance regarding the custom dataset. I have my own dataset i.e. X, Y, Z points with labels (4 columns in total) in txt files. I wanted to know how I can create a data loader for custom data and train the network.

Hi everyone,

I am pretty new in Neural Networks, can anyone share with me which parameters do you use? In which formats are (number, dictionary, array)? How do you run the class PointCNN?

Thank you in advance!

I am trying to create (and plot) a confusion matrix for a multilabel dataset that has the predicted results structured in the same way the S3DIS results. I'm having trouble with creating the confusion matrix from the results and I am wondering if anyone has figured out how to do this. Any help would be greatly appreciated!