之前有段时间搞过3d人体姿态估计,这个环境配了一下午,现在总结下,避免跳坑
conda create -n mmhuman3d python=3.8
source activate mmhuman3d
conda install ffmpeg
conda install pytorch==1.10.0 torchvision==0.11.0 cudatoolkit=11.3 -c pytorch -c conda-forge
conda install -c fvcore -c iopath -c conda-forge fvcore iopath -y
conda install -c bottler nvidiacub -y
conda install pytorch3d -c pytorch3d
cd mmhuman3d
pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
Modify the line extra_compile_args = {"cxx": ["-std=c++14"]} to extra_compile_args = {"cxx": []} in setup.py (around line 40).
Remove the "-std=c++14", line (around line 60) from nvcc_args in setup.py.
export CUDA_HOME=/usr/local/cuda-11.4
pip install -v -e .
pip install "mmcv-full>=1.3.17,<=1.5.3" -f https://download.openmmlab.com/mmcv/dist/cu113/torch1.10.0/index.html -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install "mmdet<=2.25.1" -i https://pypi.tuna.tsinghua.edu.cn/simple (demo -> must)
pip install "mmpose<=0.28.1" -i https://pypi.tuna.tsinghua.edu.cn/simple (demo -> must)
pip install "mmcls<=0.23.2" "mmtrack<=0.13.0" -i https://pypi.tuna.tsinghua.edu.cn/simple