Frequently Asked Questions#

We list some common troubles faced by many users and their corresponding solutions here. Before posting an issue, please go through the following troubleshooting steps first:

  • Check whether the issue is TorchSparse specific or environment specific. Try creating an isolated environment via Docker or on another computer and see if the error persists. If using TorchSparse as a dependancy of another project, ensure the downstream project is compatible with the version of TorchSparse that you installed.

  • Read the error logs line-by-line: if it’s a compilation error, the problem will be shown in the log. Often, compilation issues will come from incorrectly configured environment, such as an improper NVCC or PyTorch installation, rather than incompatibility with this library. Please paste the full log message of pip install -v git+https://github.com/mit-han-lab/torchsparse.git when you submit the issue.

  • Try completely uninstalling CUDA and make sure that there are no additional CUDA installations:

    ls /usr/local/cuda* -d

  • Then, follow all of the steps for toolkit installation in the CUDA installation guide, especially the post installation actions to set your LD_LIBRARY_PATH and PATH.

  • Ensure that PyTorch and NVCC are using the same version of CUDA:

    nvcc --version
python -c "import torch; print(torch.version.cuda);"

  • If you’re trying to cross-compile the library (i.e. compiling for a different GPU than the one in the system at build time, such as in a docker build), make use of the TORCH_CUDA_ARCH_LIST environmental variable. You can use this chart to find your architecture/gencode. For example, if you want to compile for a Turing-architecture GPU, you would do:

    TORCH_CUDA_ARCH_LIST="7.0;7.5" pip install --upgrade git+https://github.com/mit-han-lab/torchsparse.git

  • If you see Killed in the compilation log, it’s likely the compilation failed due to out of memory as a result of parallel compilation. You can limit the number of CPUs the compiler will use by setting the MAX_JOBS environmental variable before installation:

    MAX_JOBS=2 pip install --upgrade git+https://github.com/mit-han-lab/torchsparse.git