INSTALL.md

Installation

Code

The code has been tested with Python 3.8, CUDA 10.2 and PyTorch 1.7.1 on Ubuntu 18.04.

• Clone this repo, create virtual environment & install requirements

  git clone git@github.com:muelea/shapy.git
  cd shapy
  export PYTHONPATH=$PYTHONPATH:$(pwd)/attributes/
  
  python3.8 -m venv .venv/shapy
  source .venv/shapy/bin/activate
  pip install -r requirements.txt
  
  cd attributes
  python setup.py install
  
  cd ../mesh-mesh-intersection
  export CUDA_SAMPLES_INC=$(pwd)/include
  pip install -r requirements.txt
  python setup.py install
  

Body model and model data

Folder structure

In shapy/data, you will need subfolders for the neutral SMPL-X body model (body_models) and ExPose and SHAPY models and utilities (expose_release, trained_models, utility_files). The final data structure should look like this:

data
├── body_models
├── expose_release
├── trained_models
└── utility_files

SMPL Model

Download the neutral SMPL-X Model from the official website. You can also optionally download SMPL Your body model subfolder should have the following structure:

data
├── body_models
    └── smpl
        ├── SMPL_NEUTRAL.pkl
        ├── SMPL_FEMALE.pkl
        ├── SMPL_MALE.pkl
    └── smplx
        ├── SMPLX_NEUTRAL.npz
        ├── SMPLX_FEMALE.npz
        ├── SMPLX_MALE.npz

ExPose and SHAPY utilities

Option 1 (chose if you have not yet registered on the SHAPY website)

Download shapy_data.zip from our website and extract it in the data folder:

cd data
unzip shapy_data.zip

Option 2 (chose if you have already registered on the SHAPY website)

Run download_data.sh. This will request your username and password for the SHAPY website and then download and extract the SHAPY model data.

cd data
bash download_data.sh

Complete folder structure

After that, you should have the following structure:

data
├── body_models
│   ├── smpl
│   └── smplx
├── expose_release
│   ├── data
│   │   ├── all_means.pkl
│   │   └── SMPLX_to_J14.pkl
│   └── utility_files
│       └── flame
│           └── SMPL-X__FLAME_vertex_ids.npy
├── trained_models
│   ├── a2b
│   │   ├── caesar-female_smplx-female-10betas
│   │   ├── caesar-female_smplx-neutral-10betas
│   │   ├── caesar-male_smplx-male-10betas
│   │   └── caesar-male_smplx-neutral-10betas
│   ├── b2a
│   │   └── polynomial
│   │       ├── caesar-female_smplx-female-10betas
│   │       ├── caesar-female_smplx-neutral-10betas
│   │       ├── caesar-male_smplx-male-10betas
│   │       └── caesar-male_smplx-neutral-10betas
│   └── shapy
│       └── SHAPY_A
└── utility_files
    ├── evaluation
    │   └── eval_point_set
    │       ├── HD_SMPL_sparse.pkl
    │       └── HD_SMPLX_from_SMPL.pkl
    ├── measurements
    │   ├── measurement_defitions.yaml
    │   ├── smpl_measurement_vertices.yaml
    │   └── smplx_measurements.yaml
    ├── shape_priors
    │   ├── female_normal.npz
    │   └── male_normal.npz
    └── smplx
        ├── smplx_correspondences.npz
        └── smplx_extra_joints.yaml