cszyzhang / riconv2 Goto Github PK

View Code? Open in Web Editor NEW

25.0 2.0 3.0 37.87 MB

The official source codes for the riconv++ (IJCV2022) paper

License: MIT License

Python 85.20% C++ 4.30% Cuda 9.38% C 1.12%

deep-learning point-cloud rotation-invariant cnn classification segmentation

riconv2's People

Contributors

Stargazers

Watchers

Forkers

yingruijie rancheng dolphin4mi

riconv2's Issues

About LRA calculation

Thank you for sharing the code.

I have a question about the LRA calculation.
Why didn't you remove the sign ambiguity of the eigen vectors like SHOT feature?
Is there any reason?

Question about ModelNet40 experiment

Hi @cszyzhang ,

Thank you for sharing the code.

I want to train and test RIConv++ on ModelNet40 dataset for z/SO(3) case.
I checked in riconv2/data_utils/ModelNetDataLoader.py, but did not find any code that applies random rotation to point clouds.
Could you please tell me where in the code do you apply random z and SO(3) rotation to ModelNet40 point cloud?

Thank you,
Supriya

PyTorch 1.11 above build pointnet2 error THC not found

There might be several errors triggered by the depreciated THC library.

Emitting ninja build file /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/build.ninja...
Compiling objects...
Allowing ninja to set a default number of workers... (overridable by setting the environment variable MAX_JOBS=N)
[1/9] c++ -MMD -MF /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/ball_query.o.d -pthread -B /home/ran/miniconda3/envs/icra_gpr_s3e/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/ball_query.cpp -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/ball_query.o -g -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++14
FAILED: /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/ball_query.o 
c++ -MMD -MF /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/ball_query.o.d -pthread -B /home/ran/miniconda3/envs/icra_gpr_s3e/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/ball_query.cpp -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/ball_query.o -g -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++14
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
/home/ran/PycharmProjects/riconv2/models/pointnet2/src/ball_query.cpp:3:10: fatal error: THC/THC.h: No such file or directory
 #include <THC/THC.h>
          ^~~~~~~~~~~
compilation terminated.
[2/9] c++ -MMD -MF /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/group_points.o.d -pthread -B /home/ran/miniconda3/envs/icra_gpr_s3e/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/group_points.cpp -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/group_points.o -g -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++14
FAILED: /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/group_points.o 
c++ -MMD -MF /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/group_points.o.d -pthread -B /home/ran/miniconda3/envs/icra_gpr_s3e/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/group_points.cpp -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/group_points.o -g -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++14
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
/home/ran/PycharmProjects/riconv2/models/pointnet2/src/group_points.cpp:5:10: fatal error: THC/THC.h: No such file or directory
 #include <THC/THC.h>
          ^~~~~~~~~~~
compilation terminated.
[3/9] c++ -MMD -MF /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/interpolate.o.d -pthread -B /home/ran/miniconda3/envs/icra_gpr_s3e/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/interpolate.cpp -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/interpolate.o -g -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++14
FAILED: /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/interpolate.o 
c++ -MMD -MF /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/interpolate.o.d -pthread -B /home/ran/miniconda3/envs/icra_gpr_s3e/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/interpolate.cpp -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/interpolate.o -g -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++14
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
/home/ran/PycharmProjects/riconv2/models/pointnet2/src/interpolate.cpp:3:10: fatal error: THC/THC.h: No such file or directory
 #include <THC/THC.h>
          ^~~~~~~~~~~
compilation terminated.
[4/9] c++ -MMD -MF /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/sampling.o.d -pthread -B /home/ran/miniconda3/envs/icra_gpr_s3e/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/sampling.cpp -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/sampling.o -g -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++14
FAILED: /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/sampling.o 
c++ -MMD -MF /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/sampling.o.d -pthread -B /home/ran/miniconda3/envs/icra_gpr_s3e/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/sampling.cpp -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/sampling.o -g -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++14
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
/home/ran/PycharmProjects/riconv2/models/pointnet2/src/sampling.cpp:4:10: fatal error: THC/THC.h: No such file or directory
 #include <THC/THC.h>
          ^~~~~~~~~~~
compilation terminated.
[5/9] c++ -MMD -MF /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/pointnet2_api.o.d -pthread -B /home/ran/miniconda3/envs/icra_gpr_s3e/compiler_compat -Wl,--sysroot=/ -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/pointnet2_api.cpp -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/pointnet2_api.o -g -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -std=c++14
cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++
[6/9] /usr/local/cuda/bin/nvcc  -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/interpolate_gpu.cu -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/interpolate_gpu.o -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr --compiler-options ''"'"'-fPIC'"'"'' -O2 -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -gencode=arch=compute_86,code=compute_86 -gencode=arch=compute_86,code=sm_86 -std=c++14
[7/9] /usr/local/cuda/bin/nvcc  -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/ball_query_gpu.cu -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/ball_query_gpu.o -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr --compiler-options ''"'"'-fPIC'"'"'' -O2 -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -gencode=arch=compute_86,code=compute_86 -gencode=arch=compute_86,code=sm_86 -std=c++14
[8/9] /usr/local/cuda/bin/nvcc  -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/group_points_gpu.cu -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/group_points_gpu.o -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr --compiler-options ''"'"'-fPIC'"'"'' -O2 -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -gencode=arch=compute_86,code=compute_86 -gencode=arch=compute_86,code=sm_86 -std=c++14
[9/9] /usr/local/cuda/bin/nvcc  -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/torch/csrc/api/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/TH -I/home/ran/miniconda3/envs/icra_gpr_s3e/lib/python3.8/site-packages/torch/include/THC -I/usr/local/cuda/include -I/home/ran/miniconda3/envs/icra_gpr_s3e/include/python3.8 -c -c /home/ran/PycharmProjects/riconv2/models/pointnet2/src/sampling_gpu.cu -o /home/ran/PycharmProjects/riconv2/models/pointnet2/build/temp.linux-x86_64-3.8/src/sampling_gpu.o -D__CUDA_NO_HALF_OPERATORS__ -D__CUDA_NO_HALF_CONVERSIONS__ -D__CUDA_NO_BFLOAT16_CONVERSIONS__ -D__CUDA_NO_HALF2_OPERATORS__ --expt-relaxed-constexpr --compiler-options ''"'"'-fPIC'"'"'' -O2 -DTORCH_API_INCLUDE_EXTENSION_H '-DPYBIND11_COMPILER_TYPE="_gcc"' '-DPYBIND11_STDLIB="_libstdcpp"' '-DPYBIND11_BUILD_ABI="_cxxabi1011"' -DTORCH_EXTENSION_NAME=pointnet2_cuda -D_GLIBCXX_USE_CXX11_ABI=0 -gencode=arch=compute_86,code=compute_86 -gencode=arch=compute_86,code=sm_86 -std=c++14
ninja: build stopped: subcommand failed.

error: ‘getCurrentCUDAStream’ is not a member of ‘at::cuda’ cudaStream_t stream = at::cuda::getCurrentCUDAStream();

Question on Experiments

I am a student and recently I was learning about rotation invariance of point clouds. After reading your article, I benefited a lot, but I have a few questions about the experiments. Why doesn't any rotation transformations exist on dataset class, and what kind of rotation augmentation used on training for the pretrained models. Thanks a lot.

About shape retrieval

After reading your article, I am reproducing your method. I would like to ask when the source code of the shape retrieval part will be open-sourced？Thank you.

About the normal channel

In your paper, you provided the result when not using the normal channel. However, it leads to an error when testing the data without normal.

It seems that LRA is not computed when normal is None.

Can't predict on single batch

Hello!
First of all, congrats to the team for the approach and sharing the code!

After playing around with it and trying to test the part segmentation model on my augmented shapenet dataset I found that I cannot predict just one sample.

Steps done so far:

Installed/setup everything and test_partseg.py runs without any issue.
Recreating you script to test in custom data set (similar to shapenet):
- Loaded the model structure (similar to your get_model class in riconv2_part_seg.py)
- Successfully loaded the pretained model weights (thanks to sharing that too)
- Loaded my test sample, sampled it (2048 pts by default), and since it was an Airplane, assigned label to 0
- After applying the same preprocessing steps and unsqueeze it to add the batch dimension I thought everything was ready to go

On prediction I get:

RuntimeError                              Traceback (most recent call last)
Input In [341], in <cell line: 1>()
----> 1 seg_pred, _ = model(X, to_categorical(cat, num_classes))

File ~\Anaconda3\envs\riconv\lib\site-packages\torch\nn\modules\module.py:1051, in Module._call_impl(self, *input, **kwargs)
   1047 # If we don't have any hooks, we want to skip the rest of the logic in
   1048 # this function, and just call forward.
   1049 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
   1050         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1051     return forward_call(*input, **kwargs)
   1052 # Do not call functions when jit is used
   1053 full_backward_hooks, non_full_backward_hooks = [], []

Input In [296], in RIConv2.forward(self, xyz, cls_label)
     35     xyz = xyz[:, :, :3]
     37 l0_xyz, l0_norm, l0_points = self.sa0(xyz, norm, None)
---> 38 l1_xyz, l1_norm, l1_points = self.sa1(l0_xyz, l0_norm, l0_points)
     39 l2_xyz, l2_norm, l2_points = self.sa2(l1_xyz, l1_norm, l1_points)
     40 l3_xyz, l3_norm, l3_points = self.sa3(l2_xyz, l2_norm, l2_points)

File ~\Anaconda3\envs\riconv\lib\site-packages\torch\nn\modules\module.py:1051, in Module._call_impl(self, *input, **kwargs)
   1047 # If we don't have any hooks, we want to skip the rest of the logic in
   1048 # this function, and just call forward.
   1049 if not (self._backward_hooks or self._forward_hooks or self._forward_pre_hooks or _global_backward_hooks
   1050         or _global_forward_hooks or _global_forward_pre_hooks):
-> 1051     return forward_call(*input, **kwargs)
   1052 # Do not call functions when jit is used
   1053 full_backward_hooks, non_full_backward_hooks = [], []

File C:\workspace\projects\research\pcl\pcl.segmentation\riconv2\models\riconv2_utils.py:401, in RIConv2SetAbstraction.forward(self, xyz, norm, points)
    399     new_xyz, ri_feat, new_norm, idx_ordered = sample_and_group_all(xyz, norm)
    400 else:
--> 401     new_xyz, ri_feat, new_norm, idx_ordered = sample_and_group(self.npoint, self.radius, self.nsample, xyz,
    402                                                                norm)
    404 # lift
    405 ri_feat = ri_feat.permute(0, 3, 2, 1)

File C:\workspace\projects\research\pcl\pcl.segmentation\riconv2\models\riconv2_utils.py:321, in sample_and_group(npoint, radius, nsample, xyz, norm)
    318 xyz = xyz.contiguous()
    319 norm = norm.contiguous()
--> 321 new_xyz, new_norm = sample(npoint, xyz, norm=norm, sampling='fps')
    322 idx = group_index(nsample, radius, xyz, new_xyz, group='knn')
    324 ri_feat, idx_ordered = RI_features(xyz, norm, new_xyz, new_norm, idx)

File C:\workspace\projects\research\pcl\pcl.segmentation\riconv2\models\riconv2_utils.py:196, in sample(npoint, xyz, norm, sampling)
    194     new_xyz = index_points(xyz, fps_idx)
    195     if norm is not None:
--> 196         new_norm = index_points(norm, fps_idx)
    197 elif sampling == 'random':
    198     shuffle = np.arange(xyz.shape[1])

File C:\workspace\projects\research\pcl\pcl.segmentation\riconv2\models\riconv2_utils.py:70, in index_points(points, idx)
     68 repeat_shape = list(idx.shape)
     69 repeat_shape[0] = 1
---> 70 batch_indices = torch.arange(B, dtype=torch.long).to(device).view(view_shape).repeat(repeat_shape)
     72 new_points = points[batch_indices, idx, :]
     73 return new_points

RuntimeError: shape '[1, 1]' is invalid for input of size 512

If I load the test dataset as you do in test_partseg.py into a test dataLoader it works perfectly until... I set the batch size to 1 and try it (I then get the same RuntimeError as above).

With batchsize > 1:

Now setting batchsize = 1:

I hope you can help me with this.
Thanks in advance.

Best regards,
Raphael Espanha

No module named 'pointnet2_cuda'

Hi, thanks for your open source code firstly, but I met a problem that "No module named 'pointnet2_cuda'" even though I had run the setup.py without any errors. I 'll appreciate it if you can solve my problem.

xyz

How to train using only xyz?

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.