Run example.sh error

[neilwang0913]

Hi: After running the example.sh, I got an error for no: import BACore ModuleNotFoundError: No module named 'BACore'

Any idea to fix it?

Many thanks

[Penterakt]

Same problem here.

sh example.sh  ✔ -- The CXX compiler identification is GNU 11.3.0 -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Check for working CXX compiler: /opt/cuda/bin/g++ - skipped -- Detecting CXX compile features -- Detecting CXX compile features - done -- Found OpenMP_CXX: -fopenmp (found version "4.5") -- Found OpenMP: TRUE (found version "4.5")
-- The CUDA compiler identification is NVIDIA 11.7.99 -- Detecting CUDA compiler ABI info -- Detecting CUDA compiler ABI info - done -- Check for working CUDA compiler: /opt/cuda/bin/nvcc - skipped -- Detecting CUDA compile features -- Detecting CUDA compile features - done /usr/lib/python3.10/site-packages/torch -- pybind11 v2.6.3 dev1 CMake Warning (dev) at /var/lib/snapd/snap/cmake/1156/share/cmake-3.24/Modules/CMakeDependentOption.cmake:89 (message): Policy CMP0127 is not set: cmake_dependent_option() supports full Condition Syntax. Run "cmake --help-policy CMP0127" for policy details. Use the cmake_policy command to set the policy and suppress this warning. Call Stack (most recent call first): 3rd/pybind11/CMakeLists.txt:98 (cmake_dependent_option) This warning is for project developers. Use -Wno-dev to suppress it.

-- Found PythonInterp: /usr/bin/python3.10 (found version "3.10.6") -- Found PythonLibs: /usr/lib/libpython3.10.so -- Performing Test HAS_FLTO -- Performing Test HAS_FLTO - Success -- Configuring done CMake Warning (dev) in CMakeLists.txt: Policy CMP0104 is not set: CMAKE_CUDA_ARCHITECTURES now detected for NVCC, empty CUDA_ARCHITECTURES not allowed. Run "cmake --help-policy CMP0104" for policy details. Use the cmake_policy command to set the policy and suppress this warning.

CUDA_ARCHITECTURES is empty for target "LMCoreKernel". This warning is for project developers. Use -Wno-dev to suppress it.

CMake Warning (dev) in CMakeLists.txt: Policy CMP0104 is not set: CMAKE_CUDA_ARCHITECTURES now detected for NVCC, empty CUDA_ARCHITECTURES not allowed. Run "cmake --help-policy CMP0104" for policy details. Use the cmake_policy command to set the policy and suppress this warning.

CUDA_ARCHITECTURES is empty for target "LMCoreKernel". This warning is for project developers. Use -Wno-dev to suppress it.

-- Generating done -- Build files have been written to: /home/yunnan/repos/DeepLM/build [ 10%] Building CUDA object CMakeFiles/LMCoreKernel.dir/TorchLM/cpp/kernel_impl.cu.o [ 20%] Building CXX object CMakeFiles/BACore.dir/BAProblem/cpp/baproblem_manager.cc.o [ 30%] Building CXX object CMakeFiles/BACore.dir/BAProblem/cpp/interface.cc.o [ 40%] Building CXX object CMakeFiles/BACore.dir/BAProblem/cpp/io.cc.o [ 50%] Building CXX object CMakeFiles/BACore.dir/BAProblem/cpp/torch_util.cc.o [ 60%] Linking CUDA shared library libLMCoreKernel.so [ 60%] Built target LMCoreKernel [ 80%] Building CXX object CMakeFiles/LMCore.dir/TorchLM/cpp/kernel.cc.o [ 80%] Building CXX object CMakeFiles/LMCore.dir/TorchLM/cpp/interface.cc.o /home/yunnan/repos/DeepLM/BAProblem/cpp/baproblem_manager.cc: In function ‘std::vector<std::vector > PrepareSeparator(const at::Tensor&, const at::Tensor&, const at::Tensor&, const at::Tensor&, at::Tensor&, long int)’: /home/yunnan/repos/DeepLM/BAProblem/cpp/baproblem_manager.cc:118:28: warning: comparison of integer expressions of different signedness: ‘long int’ and ‘std::vector<std::vector >::size_type’ {aka ‘long unsigned int’} [-Wsign-compare] 118 | for (long i = 0; i < originPtIndices.size(); ++i) { | ^~~~~~ /home/yunnan/repos/DeepLM/BAProblem/cpp/baproblem_manager.cc:64:21: warning: unused variable ‘dPtIdx’ [-Wunused-variable] 64 | const long dPtIdx = static_cast<const long>(pointIdx.storage().data()); | ^~ /home/yunnan/repos/DeepLM/BAProblem/cpp/baproblem_manager.cc:117:14: warning: variable ‘intOptions’ set but not used [-Wunused-but-set-variable] 117 | auto intOptions = torch::TensorOptions().dtype(torch::kInt64); | ^~~~~~ /home/yunnan/repos/DeepLM/BAProblem/cpp/io.cc: In function ‘std::vector LoadBALFromFile(const char, int, int, int)’: /home/yunnan/repos/DeepLM/BAProblem/cpp/io.cc:44:27: warning: comparison of integer expressions of different signedness: ‘int’ and ‘std::vector<double, std::allocator >::size_type’ {aka ‘long unsigned int’} [-Wsign-compare] 44 | for (int i = 0; i < cameraParameters.size(); ++i) | ^~~~~~~ /home/yunnan/repos/DeepLM/BAProblem/cpp/io.cc:49:27: warning: comparison of integer expressions of different signedness: ‘int’ and ‘std::vector<double, std::allocator >::size_type’ {aka ‘long unsigned int’} [-Wsign-compare] 49 | for (int i = 0; i < points3d.size(); ++i) { | ^~~~~~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc: In function ‘void JacobiColumnSquare(const std::vector&, const std::vector&, std::vector&, int)’: /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:261:27: warning: comparison of integer expressions of different signedness: ‘int’ and ‘std::vector::size_type’ {aka ‘long unsigned int’} [-Wsign-compare] 261 | for (int i = 0; i < jacobians.size(); ++i) { | ^~~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:258:13: warning: unused variable ‘residualDim’ [-Wunused-variable] 258 | int residualDim = indices.size(); | ^~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc: In function ‘void ColumnInverseSquare(std::vector&)’: /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:301:27: warning: comparison of integer expressions of different signedness: ‘int’ and ‘std::vector::size_type’ {aka ‘long unsigned int’} [-Wsign-compare] 301 | for (int i = 0; i < jacobianScale.size(); ++i) { | ^~~~~~~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc: In function ‘void JacobiNormalize(const std::vector&, const std::vector&, std::vector&)’: /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:327:27: warning: comparison of integer expressions of different signedness: ‘int’ and ‘std::vector::size_type’ {aka ‘long unsigned int’} [-Wsign-compare] 327 | for (int i = 0; i < jacobianScale.size(); ++i) { | ^~~~~~~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:335:21: warning: unused variable ‘num’ [-Wunused-variable] 335 | int num = numDimV numDimP; | ^~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:325:13: warning: unused variable ‘residualDim’ [-Wunused-variable] 325 | int residualDim = indices.size(); | ^~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc: In function ‘void JacobiLeftMultiplyCuda(const std::vector&, const at::Tensor&, const std::vector&, const at::Tensor&, std::vector&, int)’: /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:408:21: warning: unused variable ‘numDimV’ [-Wunused-variable] 408 | int numDimV = jtr.size(0); | ^~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc: In function ‘void JacobiColumnSquareCuda(const std::vector&, const std::vector&, std::vector&, int)’: /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:481:27: warning: comparison of integer expressions of different signedness: ‘int’ and ‘std::vector::size_type’ {aka ‘long unsigned int’} [-Wsign-compare] 481 | for (int i = 0; i < jacobians.size(); ++i) { | ^~~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc: In function ‘void ColumnInverseSquareCuda(std::vector&)’: /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:499:27: warning: comparison of integer expressions of different signedness: ‘int’ and ‘std::vector::size_type’ {aka ‘long unsigned int’} [-Wsign-compare] 499 | for (int i = 0; i < jacobianScale.size(); ++i) { | ^~~~~~~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc: In function ‘void JacobiNormalizeCuda(const std::vector&, const std::vector&, std::vector&)’: /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:516:27: warning: comparison of integer expressions of different signedness: ‘int’ and ‘std::vector::size_type’ {aka ‘long unsigned int’} [-Wsign-compare] 516 | for (int i = 0; i < jacobianScale.size(); ++i) { | ^~~~~~~~ /home/yunnan/repos/DeepLM/TorchLM/cpp/kernel.cc:524:21: warning: unused variable ‘num’ [-Wunused-variable] 524 | int num = numDimV * numDimP; | ^~~ [ 90%] Linking CXX shared library BACore.cpython-310-x86_64-linux-gnu.so [ 90%] Built target BACore [100%] Linking CXX shared library LMCore.cpython-310-x86_64-linux-gnu.so [100%] Built target LMCore --2022-10-05 19:54:22-- https://grail.cs.washington.edu/projects/bal/data/ladybug/problem-49-7776-pre.txt.bz2 Loaded CA certificate '/etc/ssl/certs/ca-certificates.crt' Resolving grail.cs.washington.edu (grail.cs.washington.edu)... 2607:4000:200:14::5d, 128.208.5.93 Connecting to grail.cs.washington.edu (grail.cs.washington.edu)|2607:4000:200:14::5d|:443... connected. HTTP request sent, awaiting response... 200 OK Length: 448484 (438K) [application/x-bzip2] Saving to: ‘problem-49-7776-pre.txt.bz2’

problem-49-7776-pre 100%[===================>] 437.97K 211KB/s in 2.1s

2022-10-05 19:54:26 (211 KB/s) - ‘problem-49-7776-pre.txt.bz2’ saved [448484/448484]

Traceback (most recent call last): File "/home/yunnan/repos/DeepLM/examples/BundleAdjuster/bundle_adjuster.py", line 5, in import BACore ImportError: /home/yunnan/repos/DeepLM/build/BACore.cpython-310-x86_64-linux-gnu.so: undefined symbol: _ZN3c106detail23torchInternalAssertFailEPKcS2_jS2_RKSs

[hjwdzh]

I did not run into this error. I googled it and some one offers the solution as adding "--use-cxx11-abi", with a reference link: https://pytorch.org/TensorRT/tutorials/installation.html#installation (choosing the right ABI).

Let me know if it helps :)

[Penterakt]

In the end I installed conda environment from https://github.com/CompVis/stable-diffusion, and then it magically worked.

[dennisushi]

I had to export the following:

export TCNN_CUDA_ARCHITECTURE=86
export CUDA_HOME="/usr/local/cuda-11.7"
export PATH="/usr/local/cuda-11.7/bin:$PATH"
export LD_LIBRARY_PATH="/usr/local/cuda-11.7/lib64:$LD_LIBRARY_PATH"

before example.sh

It solved the

import BACore
ModuleNotFoundError: No module named 'BACore'

problem, but then I got

Traceback (most recent call last):
  File "examples/BundleAdjuster/bundle_adjuster.py", line 8, in <module>
    from BAProblem.rotation import AngleAxisRotatePoint
ModuleNotFoundError: No module named 'BAProblem'

Dirty fix: add sys.path.append(os.path.dirname(os.path.dirname(__file__))) to the example script.

[jucamohedano]

Hi! I'm also running into an error when running the example script, and it's different from the errors in previous issues. Any suggestions of what's wrong would be very much appreciated.

TORCH_USE_RTLD_GLOBAL=YES python3 examples/BundleAdjuster/bundle_adjuster.py --balFile ./data/problem-49-7776-pre.txt --device cuda
Load observation 31000 of 31843...       
Initial cost = 8.509125E+05, Memory = 2.321921E-03 G
Traceback (most recent call last):
  File "examples/BundleAdjuster/bundle_adjuster.py", line 40, in <module>
    numSuccessIterations = 15)
  File "/home/juanmohedano/OnePose_Plus_Plus/submodules/DeepLM/TorchLM/solver.py", line 726, in Solve
    solver.Solve()
  File "/home/juanmohedano/OnePose_Plus_Plus/submodules/DeepLM/TorchLM/solver.py", line 617, in Solve
    self.ComputeTrustRegionStep()
  File "/home/juanmohedano/OnePose_Plus_Plus/submodules/DeepLM/TorchLM/solver.py", line 560, in ComputeTrustRegionStep
    step = self.LinearSolve(lmDiagonal);
  File "/home/juanmohedano/OnePose_Plus_Plus/submodules/DeepLM/TorchLM/solver.py", line 428, in LinearSolve
    ListInvert(preconditioner)
  File "/home/juanmohedano/OnePose_Plus_Plus/submodules/DeepLM/TorchLM/listvec.py", line 25, in ListInvert
    listvec[i] = torch.inverse(listvec[i])
torch._C._LinAlgError: linalg.inv: (Batch element 0): The diagonal element 1 is zero, the inversion could not be completed because the input matrix is singular.

[brian2lee]

I did not run into this error. I googled it and some one offers the solution as adding "--use-cxx11-abi", with a reference link: https://pytorch.org/TensorRT/tutorials/installation.html#installation (choosing the right ABI).

Let me know if it helps :)

&

I had to export the following:

export TCNN_CUDA_ARCHITECTURE=86
export CUDA_HOME="/usr/local/cuda-11.7"
export PATH="/usr/local/cuda-11.7/bin:$PATH"
export LD_LIBRARY_PATH="/usr/local/cuda-11.7/lib64:$LD_LIBRARY_PATH"

before example.sh

It solved the

import BACore
ModuleNotFoundError: No module named 'BACore'

problem, but then I got

Traceback (most recent call last):
  File "examples/BundleAdjuster/bundle_adjuster.py", line 8, in <module>
    from BAProblem.rotation import AngleAxisRotatePoint
ModuleNotFoundError: No module named 'BAProblem'

Dirty fix: add sys.path.append(os.path.dirname(os.path.dirname(__file__))) to the example script.

Does not work for me, any additional solving method? I ran into so many issues building oneposeplus

[zym-njust]

I ran into the same error. I try to install BACore package but I can't find the source.

[zym-njust]

I finally found that I didn't download the 3rd party file 'eigen' and 'pybind11', which is not included in this ZIP. You have to download it manually and it works.

[VVinter-melon]

Dirty fix: add `sys.path.append(os.path.dirname(os.path.dirname(__file__)))` to the example script.

This helped, by changing the last line of example.sh to : TORCH_USE_RTLD_GLOBAL=YES python3 ../DeepLM/examples/BundleAdjuster/bundle_adjuster.py --balFile ./data/problem-49-7776-pre.txt --device cuda and adding sys.path.append(os.path.dirname(os.path.dirname(os.path.dirname(__file__)))) to the bundle_adjuster.py file

[shivakarnati]

None of the answers worked for me. from BAProblem.rotation import AngleAxisRotatePoint getting this error. Did anyone solve this?

[dadwadw233]

I met the same proplem and i finally solved it by setting the python version correctly which can be edit in the CMakeLists.txt like this 👇

set(PYTHOH3_VERSION 3.9m)

after setting the python version to 3.9 and run the command sh example.sh, i find the correct python lib was compiled successfully named 'BACore.cpython-39-x86_64-linux-gnu.so' (which is BACore.cpython-38-x86_64-linux-gnu.so before) and this solve the problem of ImportError: xxx. @shivakarnati

[dadwadw233]

i guess @Penterakt can solve the problem by installing SD's conda env mainly because the python version of it is 3.8.5 , which is compatible for the py38 version xxx.so

[titanior]

Please make sure the right cuda version is installed