Installation

June 19, 2025 ยท View on GitHub

Local paths setup

Create system.json file in the root of this directory with your machine's file paths using system.template.json as a template. system.json should include the following:

  • Path for creating logs for one-off scripts ('output')

  • Path to the folder with (generated) datasets of sewing patterns ('datasets_path')

  • Path to the folder with simulation results on the datasets of sewing patterns ('datasets_sim')

  • Data generation & Simulation resources

    • path to folder with simulation\rendering configurations ('sim_configs_path')
    • path to folder containing body files for neutral body and other base body models ('bodies_default_path')
    • path to folder containing datasets of body shape samples ('body_samples_path')

Installing simulator

We use our own version of the NVIDIA warp simulator. It should be installed manually to use our library correctly.

See the instructions in the NvidiaWarp-GarmentCode repo.

Using pip

With pip, you can install the core pygarment library and its dependencies to start writing your own garment programs!

pip install pygarment

Manual Installation

If required, you could install the library and dependecies manually

pygarment uses the following dependencies:

These dependencies are installed through setup.cfg.

All python dependencies can be installed with pip install / conda install:

conda create -n garmentcode python=3.9
conda activate garmentcode
pip install -e . # installs in editable mode, installs requirements from setup.cfg
<build and install warp for GarmentCode>

Add the root repository to PYTHONPATH.

=> The code is ready to run

NOTE: check out a full environment setup and running process from our early adopter: https://github.com/maria-korosteleva/GarmentCode/issues/17