Installation

Note

It is recommended to use mamba/conda to create a clean environment.

# python >=3.9
micromamba create -n tace python=3.13.11 -y
micromamba activate tace

You can install the package in ways as described below.

Install from Source (recommended)

git clone https://github.com/xvzemin/tace.git
cd tace
pip install .

# Optional packages for accelerating computations, highly recommended.

# openequivariance
pip install openequivariance

# cuequivariance, cuda12
pip install cuequivariance cuequivariance-torch cuequivariance-ops-torch-cu12

# cuequivariance, cuda13
pip install cuequivariance cuequivariance-torch cuequivariance-ops-torch-cu12

# equitorch
# If you want to accelerate correlation > 2 in the product basis,
# you need to install torch_scatter according to your PyTorch version.
pip install torch_scatter -f https://data.pyg.org/whl/torch-2.11.0+cu130.html

Install via pip (not recommended)

pip install tace

# Optional packages for accelerating computations, highly recommended.

# openequivariance
pip install openequivariance

# cuequivariance, cuda12
pip install cuequivariance cuequivariance-torch cuequivariance-ops-torch-cu12

# cuequivariance, cuda13
pip install cuequivariance cuequivariance-torch cuequivariance-ops-torch-cu12

# equitorch
# If you want to accelerate correlation > 2 in the product basis,
# you need to install torch_scatter according to your PyTorch version.
pip install torch_scatter -f https://data.pyg.org/whl/torch-2.11.0+cu130.html