--- smplx (custom)

Environment Setup

Simplified Environments for PALM-only

If you only want to use the data and want to use our starter scripts to visualize them, follow the simplified setup instructions below:

conda create -n palm python=3.10
conda activate palm
cd code
pip install -r requirements_palm.txt

Download MANO_RIGHT.pkl from here by registering an account. Then put the pickle in this path:

code/data/body_models/MANO_RIGHT.pkl

Full Environment Setup

General Requirements (NVIDIA A100):

• PyTorch: 2.0.1+cu118
• Pytorch 3D: 0.7.4
• Kaolin: 0.17.0
• PyTorch Lightning: 1.9.5
• Python: 3.10.14
• CUDA: 11.8 (check nvcc --version)

These requirements are non-strict. However, they have been tested on our end. Therefore, it is a good starting point.

Preliminary

PyTorch Lightning: To avoid boilerplate code, we use pytorch lightning (PL) to handle the main logic for training and evaluation. Feel free to consult the documentation, should you have any questions.

Before starting, check your CUDA nvcc version:

nvcc --version # should be 11.8

You can install nvcc and cuda via runfile. If nvcc --version is still not 11.8, check whether you are referring the right nvcc with which nvcc. Assuming you have an NVIDIA driver installed, usually, you only need to run the following command to install nvcc (as an example):

sudo bash cuda_*_linux.run --toolkit --silent --override

After the installation, make sure the paths pointing to the current cuda toolkit location. For example:

export CUDA_HOME=/usr/local/cuda-11.8
export PATH="/usr/local/cuda-11.8/bin:$PATH"
export CPATH="/usr/local/cuda-11.8/include:$CPATH"
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda-11.8/lib64/"

PALM-Net environment

# -- conda
ENV_NAME=palmnet
conda create -n $ENV_NAME python=3.10.14
conda activate $ENV_NAME

cd code

# -- dependencies
pip install -r requirements.txt
pip install pytorch-lightning==1.9.5
conda install -c conda-forge openmesh
conda install -c fvcore -c iopath -c conda-forge fvcore iopath
conda install -c bottler nvidiacub

# -- torch
pip install torch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 --index-url https://download.pytorch.org/whl/cu118

# -- submodules
mkdir submodules && cd submodules

# --- pytroch3d
git clone https://github.com/facebookresearch/pytorch3d.git 
cd pytorch3d
git checkout v0.7.4
export CC=g++-11 # if your GCC has too high version for the next line
export CXX=g++-11 # if your GCC has too high version for the next line
python setup.py install
cd ..

# --- kaolin
pip install kaolin==0.17.0 -f https://nvidia-kaolin.s3.us-east-2.amazonaws.com/torch-2.0.1_cu118.html

# --- smplx (custom)
git clone https://github.com/zc-alexfan/smplx.git
cd smplx
git checkout 0d7747f
python setup.py install
cd ../..

# --- tinycuda-nn
pip install --no-build-isolation git+https://github.com/NVlabs/tiny-cuda-nn/#subdirectory=bindings/torch

# override numpy
pip install numpy==1.23.5

Setup files

Download MANO_RIGHT.pkl from here by registering an account. Then put the pickle in this path:

code/data/body_models/MANO_RIGHT.pkl

Download our pretrained checkpoint (flat_hand.right.dim32.7subj.feb.23.pt) and put it here:

code/data/pretrained/flat_hand.right.dim32.7subj.feb.23.pt