#Topic
Running AlphaPose on WSL
#Background
Most of the monocular 3D human pose estimation (M3DHPE) models I looked at needed more than 8 GB to produce 3D pose directly from RGB.
I looked for an alternative that would first extract a 2D pose, then lift it into 3D.
| Contents | Description |
|---|---|
| Windows | Setting up multiple CUDA versions on Windows |
#Windows
#Environment
- Windows 11
- WSL 2.0
#Code
Here are the original installation commands.
github
Paper
#Commands to run
# Original AlphaPose installation method using pip
# 1. Install PyTorch
pip3 install torch torchvision --extra-index-url https://download.pytorch.org/whl/cu12.6
# Check torch environment by: python3 -m torch.utils.collect_env
# 2. Get AlphaPose
git clone https://github.com/MVIG-SJTU/AlphaPose.git
cd AlphaPose
export PATH=/usr/local/cuda/bin/:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64/:$LD_LIBRARY_PATH
pip install cython
sudo apt-get install libyaml-dev
# 4. Continue as shown
python -m pip install cython
pip install ninja easydict matplotlib opencv-python tensorboardx terminaltables visdom numpy scipy pyyaml cython_bbox
sudo apt-get install libyaml-dev
pip install --upgrade pip setuptools wheel
pip3 install git+https://github.com/Ambrosiussen/HalpeCOCOAPI.git#subdirectory=PythonAPI
python setup.py build develop
# 4. Install PyTorch3D (Optional, only for visualization)
conda install -c fvcore -c iopath -c conda-forge fvcore iopath
conda install -c bottler nvidiacub
pip install git+ssh://git@github.com/facebookresearch/pytorch3d.git@stable
# Create Pyenv virtual environment
pyenv install 3.9.21
pyenv virtualenv 3.9.21 alphapose
pyenv activate alphapose
# Install PyTorch libraries