- conda create -n pytorch3d python=3.9
- conda activate pytorch3d
- conda install -c pytorch pytorch=1.9.1 torchvision cudatoolkit=10.2
- conda install -c fvcore -c iopath -c conda-forge fvcore iopath
- Otherwise download the CUB library from https://github.com/NVIDIA/cub/releases and unpack it to a folder of your choice. Define the environment variable CUB_HOME before building and point it to the directory that contains CMakeLists.txt for CUB.
- pip install "git+https://github.com/facebookresearch/pytorch3d.git"
저는 서영준 입니다.