$ make -j

Tested on

Create a virtual environment for the required packages. We suggest conda and provide the requirements in environment.yaml.

Once the environment is created, run the following command to install the local resources.

$ conda env create -f environment.yaml
$ conda activate sigsvgd
$ pip install -e .
$ pip install pytorch_lightning==2.2.0 --no-deps

NOTE: In macOs, the following extra steps are required to install signatory.

export PATH="/opt/homebrew/opt/llvm/bin:$PATH"
export LDFLAGS="-L/opt/homebrew/opt/llvm/lib"
export CPPFLAGS="-I/opt/homebrew/opt/llvm/include"
MACOSX_DEPLOYMENT_TARGET=11.3.1 CC=clang CXX=clang++ pip install signatory==1.2.6.1.9.0 --no-binary signatory

To run some of the examples, you'll also need to install pybullet-planning. Please refer to their repository for installation instructions.

• Rebuttal experimental seed are located at https://github.com/lubaroli/stein_mpc/releases/tag/rebuttal-experiment-result