Installation

We strongly recommend using the Conda package manager to create an isolated environment. This simplifies the management of complex dependencies like PyTorch and Pinocchio (for testing).

First, clone the repository:

git clone https://github.com/YueWang996/bard.git
cd bard

Next, follow the instructions for your desired compute device.

CPU-Only Installation

If you do not have an NVIDIA GPU or do not require GPU acceleration, you can install the CPU-only version with a single command. This will automatically install a compatible, CPU-only version of PyTorch.

pip install -e ".[cpu]"

Upgrading from CPU to GPU

If you initially installed the CPU-only version and want to switch to the CUDA version, you must manually reinstall PyTorch.

  1. Uninstall the existing CPU-only PyTorch version:

    pip uninstall -y torch torchvision torchaudio

  2. Install the correct CUDA-enabled PyTorch build from the official site.

  3. Reinstall bard to ensure all dependencies are correctly linked.

    pip install -e . --force-reinstall

Developer and Test Setup

The library is rigorously tested against pinocchio to ensure numerical accuracy. To run the full test suite, you will need to install the development dependencies.

# First, install pinocchio
conda install -c conda-forge pinocchio

# From the project root directory, install dev dependencies
pip install -e ".[dev]"

# Run the test suite
pytest

Building the Documentation

If you need to build the documentation locally, follow these steps.

  1. Install documentation dependencies from the project’s root directory:

    pip install -e .[docs]

  2. Navigate to the `docs` folder and run the build command:

    cd docs
sphinx-build -b html source build/html

  3. View the documentation by opening the index.html file located in the docs/build/html/ directory in your web browser.